Artificial Analysis · 文生图原榜
AA-Image-T2I v2.1 Overall公开原榜快照，文生图，原始Elo与名次独立保留，不与其他版本、音频模式或内网AI试评合分。原榜评审者范围：validated private evaluators。比较次数不是人数。原始评测截止/分数更新时间未知；保存时间仅为快照观察时间。仅审核资料转录，不代表FLABS人工体验、质量验收或上游API身份验证。
当前原榜
AA-Image-T2I v2.1 Overall
已核对模型
105
资料快照 · 非原评测截止日
2026年10月22日 10:03
保留本协议的原始名次、Elo、比较次数和区间；不转换成0–100分，不与其他任务、音频模式或评测版本混算。价格为来源报价，比较次数不是参与者人数。原评测截止日未提供。
模型与原始评测记录
#1GPT Image 2.5 Sunburst (max)1221.96 Elo13,014次比较
- 原榜Elo
- 1221.96 Elo
- 原榜名次
- 1 名
- 比较次数（非人数）
- 13014 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1211.96 Elo
- Elo95%上界
- 1231.96 Elo
- 原榜名次区间下界
- 1 名
- 原榜名次区间上界
- 1 名
- 参考价格
- 210.72 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Generated using the OpenAI Images API with `model_id: "gpt-image-2.5-sunburst-2026-09-08"` and `quality: "max"`. Prompts were pre-pended with "Create the following as a square image:" for Text to Image, and "Edit the following to create a square image:" for Image Editing. We found without pre-pending these instructions, images would not be produced natively as squares.
模型详情 →#2GPT Image 2.5 Flare (max)1206.72 Elo12,364次比较
- 原榜Elo
- 1206.72 Elo
- 原榜名次
- 2 名
- 比较次数（非人数）
- 12364 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1196.72 Elo
- Elo95%上界
- 1216.72 Elo
- 原榜名次区间下界
- 2 名
- 原榜名次区间上界
- 2 名
- 参考价格
- 210.72 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Generated using the OpenAI Images API with `model_id: "gpt-image-2.5-flare-2026-09-08"` and `quality: "max"`. Prompts were pre-pended with "Create the following as a square image:" for Text to Image, and "Edit the following to create a square image:" for Image Editing. We found without pre-pending these instructions, images would not be produced natively as squares.
模型详情 →#3GPT Image 2 (high)1185.74 Elo12,835次比较
- 原榜Elo
- 1185.74 Elo
- 原榜名次
- 3 名
- 比较次数（非人数）
- 12835 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1174.74 Elo
- Elo95%上界
- 1196.74 Elo
- 原榜名次区间下界
- 3 名
- 原榜名次区间上界
- 4 名
- 参考价格
- 211 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Generated using the OpenAI Images API with `model_id: "gpt-image-2-2026-04-21"` and `quality: "high"`. Prompts were pre-pended with "Create the following as a square image:" for Text to Image, and "Edit the following to create a square image:" for Image Editing. We found without pre-pending these instructions, images would not be produced natively as squares.
模型详情 →#4Nano Banana 2.11175.56 Elo6,210次比较
- 原榜Elo
- 1175.56 Elo
- 原榜名次
- 4 名
- 比较次数（非人数）
- 6210 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1163.56 Elo
- Elo95%上界
- 1187.56 Elo
- 原榜名次区间下界
- 3 名
- 原榜名次区间上界
- 6 名
- 参考价格
- 33.6 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#5MAI-Image-2.61165.89 Elo8,342次比较
- 原榜Elo
- 1165.89 Elo
- 原榜名次
- 5 名
- 比较次数（非人数）
- 8342 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1153.89 Elo
- Elo95%上界
- 1177.89 Elo
- 原榜名次区间下界
- 4 名
- 原榜名次区间上界
- 6 名
- 参考价格
- 38.9 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#6Grok Imagine Image 2.01165.59 Elo7,481次比较
- 原榜Elo
- 1165.59 Elo
- 原榜名次
- 6 名
- 比较次数（非人数）
- 7481 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1153.59 Elo
- Elo95%上界
- 1177.59 Elo
- 原榜名次区间下界
- 4 名
- 原榜名次区间上界
- 6 名
- 参考价格
- 60 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Served at quality medium and 2k resolution; the API default quality (auto) serves low for generation.
模型详情 →#7Nano Banana 2 (Gemini 3.1 Flash Image)1139.99 Elo16,075次比较
- 原榜Elo
- 1139.99 Elo
- 原榜名次
- 7 名
- 比较次数（非人数）
- 16075 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1129.99 Elo
- Elo95%上界
- 1149.99 Elo
- 原榜名次区间下界
- 7 名
- 原榜名次区间上界
- 8 名
- 参考价格
- 67 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#8Muse Image1127.25 Elo8,479次比较
- 原榜Elo
- 1127.25 Elo
- 原榜名次
- 8 名
- 比较次数（非人数）
- 8479 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1116.25 Elo
- Elo95%上界
- 1138.25 Elo
- 原榜名次区间下界
- 7 名
- 原榜名次区间上界
- 12 名
- 参考价格
- 10 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#9GPT Image 1.5 (high)1122.60 Elo12,830次比较
- 原榜Elo
- 1122.6 Elo
- 原榜名次
- 9 名
- 比较次数（非人数）
- 12830 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1111.6 Elo
- Elo95%上界
- 1133.6 Elo
- 原榜名次区间下界
- 8 名
- 原榜名次区间上界
- 14 名
- 参考价格
- 133 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Generated using the OpenAI Images API with `model_id: "gpt-image-1.5"` and `quality: "high"`. Prompts were pre-pended with "Create the following as a square image:" for Text to Image, and "Edit the following to create a square image:" for Image Editing. We found without pre-pending these instructions, images would not be produced natively as squares.
模型详情 →#10MAI-Image-2.6-Flash1119.12 Elo5,482次比较
- 原榜Elo
- 1119.12 Elo
- 原榜名次
- 10 名
- 比较次数（非人数）
- 5482 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1107.12 Elo
- Elo95%上界
- 1131.12 Elo
- 原榜名次区间下界
- 8 名
- 原榜名次区间上界
- 15 名
- 参考价格
- 19.5 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#11Nano Banana Pro (Gemini 3 Pro Image)1117.94 Elo15,174次比较
- 原榜Elo
- 1117.94 Elo
- 原榜名次
- 11 名
- 比较次数（非人数）
- 15174 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1107.94 Elo
- Elo95%上界
- 1127.94 Elo
- 原榜名次区间下界
- 8 名
- 原榜名次区间上界
- 14 名
- 参考价格
- 134 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#12MAI-Image-2.51117.94 Elo13,789次比较
- 原榜Elo
- 1117.94 Elo
- 原榜名次
- 12 名
- 比较次数（非人数）
- 13789 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1106.94 Elo
- Elo95%上界
- 1128.94 Elo
- 原榜名次区间下界
- 8 名
- 原榜名次区间上界
- 14 名
- 参考价格
- 48.128 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#13FLUX 3 Image1117.15 Elo6,115次比较
- 原榜Elo
- 1117.15 Elo
- 原榜名次
- 13 名
- 比较次数（非人数）
- 6115 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1105.15 Elo
- Elo95%上界
- 1129.15 Elo
- 原榜名次区间下界
- 8 名
- 原榜名次区间上界
- 15 名
- 参考价格
- 48 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#14MAI-Image-2.5-Pro1114.05 Elo12,200次比较
- 原榜Elo
- 1114.05 Elo
- 原榜名次
- 14 名
- 比较次数（非人数）
- 12200 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1104.05 Elo
- Elo95%上界
- 1124.05 Elo
- 原榜名次区间下界
- 10 名
- 原榜名次区间上界
- 15 名
- 参考价格
- 108.544 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#15Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)1106.69 Elo15,080次比较
- 原榜Elo
- 1106.69 Elo
- 原榜名次
- 15 名
- 比较次数（非人数）
- 15080 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1096.69 Elo
- Elo95%上界
- 1116.69 Elo
- 原榜名次区间下界
- 12 名
- 原榜名次区间上界
- 16 名
- 参考价格
- 33.6 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#16Seedream 5.0 Pro1100.17 Elo11,633次比较
- 原榜Elo
- 1100.17 Elo
- 原榜名次
- 16 名
- 比较次数（非人数）
- 11633 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1089.17 Elo
- Elo95%上界
- 1111.17 Elo
- 原榜名次区间下界
- 15 名
- 原榜名次区间上界
- 17 名
- 参考价格
- 90 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#17Qwen-Image-3.0-Pro1095.10 Elo6,332次比较
- 原榜Elo
- 1095.1 Elo
- 原榜名次
- 17 名
- 比较次数（非人数）
- 6332 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1083.1 Elo
- Elo95%上界
- 1107.1 Elo
- 原榜名次区间下界
- 15 名
- 原榜名次区间上界
- 17 名
- 参考价格
- 40 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#18Qwen-Image-3.01070.42 Elo6,090次比较
- 原榜Elo
- 1070.42 Elo
- 原榜名次
- 18 名
- 比较次数（非人数）
- 6090 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1058.42 Elo
- Elo95%上界
- 1082.42 Elo
- 原榜名次区间下界
- 18 名
- 原榜名次区间上界
- 19 名
- 参考价格
- 30 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#19Qwen-Image-2.1-Pro1056.45 Elo4,970次比较
- 原榜Elo
- 1056.45 Elo
- 原榜名次
- 19 名
- 比较次数（非人数）
- 4970 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1044.45 Elo
- Elo95%上界
- 1068.45 Elo
- 原榜名次区间下界
- 18 名
- 原榜名次区间上界
- 21 名
- 参考价格
- 35.333 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Generated through the Alibaba Cloud Model Studio API (`qwen-image-2.1-pro`) at 2048x2048, with prompt enhancement enabled at the "direct" setting. The open weights model, Qwen-Image-2.1, runs on Artificial Analysis hardware and is listed separately.
模型详情 →#20grok-imagine-image-quality1048.86 Elo16,728次比较
- 原榜Elo
- 1048.86 Elo
- 原榜名次
- 20 名
- 比较次数（非人数）
- 16728 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1039.86 Elo
- Elo95%上界
- 1057.86 Elo
- 原榜名次区间下界
- 19 名
- 原榜名次区间上界
- 21 名
- 参考价格
- 50 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#21Qwen-Image-2.11042.13 Elo5,712次比较
- 原榜Elo
- 1042.13 Elo
- 原榜名次
- 21 名
- 比较次数（非人数）
- 5712 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 1030.13 Elo
- Elo95%上界
- 1054.13 Elo
- 原榜名次区间下界
- 20 名
- 原榜名次区间上界
- 25 名
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 本源本次显示“No API available”，报价待核验；未采用隐藏raw价unknown，这不证明发布方没有任何API。 来源说明：Open weights (`Qwen/Qwen-Image-2.1`) run on Artificial Analysis hardware at 1024x1024, with prompt enhancement enabled using Qwen's official prompt enhancer (`Qwen/Qwen-Image-2.1-PE-T2I`). Alibaba's API models, Qwen-Image-2.1-Pro and Qwen-Image-2.1-Turbo, are listed separately.
模型详情 →#22Fibo 2.01038.79 Elo5,113次比较
- 原榜Elo
- 1038.79 Elo
- 原榜名次
- 22 名
- 比较次数（非人数）
- 5113 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 1026.79 Elo
- Elo95%上界
- 1050.79 Elo
- 原榜名次区间下界
- 20 名
- 原榜名次区间上界
- 28 名
- 参考价格
- 32 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#23Krea 2 Large1033.78 Elo14,513次比较
- 原榜Elo
- 1033.78 Elo
- 原榜名次
- 23 名
- 比较次数（非人数）
- 14513 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1024.78 Elo
- Elo95%上界
- 1042.78 Elo
- 原榜名次区间下界
- 21 名
- 原榜名次区间上界
- 29 名
- 参考价格
- 60 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#24Qwen Image 2.0 Pro (2026-04-22)1029.57 Elo6,138次比较
- 原榜Elo
- 1029.57 Elo
- 原榜名次
- 24 名
- 比较次数（非人数）
- 6138 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1018.5699999999999 Elo
- Elo95%上界
- 1040.57 Elo
- 原榜名次区间下界
- 22 名
- 原榜名次区间上界
- 35 名
- 参考价格
- 75 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#25MAI-Image-2.5-Flash1028.74 Elo5,855次比较
- 原榜Elo
- 1028.74 Elo
- 原榜名次
- 25 名
- 比较次数（非人数）
- 5855 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1017.74 Elo
- Elo95%上界
- 1039.74 Elo
- 原榜名次区间下界
- 22 名
- 原榜名次区间上界
- 36 名
- 参考价格
- 19.968 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#26Qwen-Image-2.1-Turbo1028.57 Elo4,625次比较
- 原榜Elo
- 1028.57 Elo
- 原榜名次
- 26 名
- 比较次数（非人数）
- 4625 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1016.5699999999999 Elo
- Elo95%上界
- 1040.57 Elo
- 原榜名次区间下界
- 22 名
- 原榜名次区间上界
- 37 名
- 参考价格
- 16 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Generated through the Alibaba Cloud Model Studio API (`qwen-image-2.1-turbo`) at 1024x1024, with prompt enhancement enabled at the "direct" setting. The open weights model, Qwen-Image-2.1, runs on Artificial Analysis hardware and is listed separately.
模型详情 →#27Wan 2.6 Image1026.54 Elo5,370次比较
- 原榜Elo
- 1026.54 Elo
- 原榜名次
- 27 名
- 比较次数（非人数）
- 5370 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1014.54 Elo
- Elo95%上界
- 1038.54 Elo
- 原榜名次区间下界
- 22 名
- 原榜名次区间上界
- 38 名
- 参考价格
- 30 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#28Recraft V4.1 Utility1026.46 Elo7,452次比较
- 原榜Elo
- 1026.46 Elo
- 原榜名次
- 28 名
- 比较次数（非人数）
- 7452 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1015.46 Elo
- Elo95%上界
- 1037.46 Elo
- 原榜名次区间下界
- 23 名
- 原榜名次区间上界
- 37 名
- 参考价格
- 35 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#29Seedream 4.51026.34 Elo6,392次比较
- 原榜Elo
- 1026.34 Elo
- 原榜名次
- 29 名
- 比较次数（非人数）
- 6392 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1014.3399999999999 Elo
- Elo95%上界
- 1038.34 Elo
- 原榜名次区间下界
- 23 名
- 原榜名次区间上界
- 38 名
- 参考价格
- 40 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#30FLUX.2 [flex]1026.19 Elo6,082次比较
- 原榜Elo
- 1026.19 Elo
- 原榜名次
- 30 名
- 比较次数（非人数）
- 6082 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1015.19 Elo
- Elo95%上界
- 1037.19 Elo
- 原榜名次区间下界
- 23 名
- 原榜名次区间上界
- 38 名
- 参考价格
- 60 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Generated with the default settings suggested on the Black Forest Labs API `{"steps": 50, "guidance_scale": 4.5}`
模型详情 →#31Ideogram 4.0 (Quality)1025.98 Elo10,688次比较
- 原榜Elo
- 1025.98 Elo
- 原榜名次
- 31 名
- 比较次数（非人数）
- 10688 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 1015.98 Elo
- Elo95%上界
- 1035.98 Elo
- 原榜名次区间下界
- 23 名
- 原榜名次区间上界
- 37 名
- 参考价格
- 100 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Generated with Ideogram's first party API for Ideogram 4.0, which is different from the publicly available open weights model that has additional safety training and quantization.
模型详情 →#32Luma UNI 1 Max1023.21 Elo5,539次比较
- 原榜Elo
- 1023.21 Elo
- 原榜名次
- 32 名
- 比较次数（非人数）
- 5539 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1011.21 Elo
- Elo95%上界
- 1035.21 Elo
- 原榜名次区间下界
- 24 名
- 原榜名次区间上界
- 40 名
- 参考价格
- 100 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#33grok-imagine-image1022.23 Elo6,330次比较
- 原榜Elo
- 1022.23 Elo
- 原榜名次
- 33 名
- 比较次数（非人数）
- 6330 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1011.23 Elo
- Elo95%上界
- 1033.23 Elo
- 原榜名次区间下界
- 24 名
- 原榜名次区间上界
- 40 名
- 参考价格
- 20 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#34FLUX.2 [max]1021.16 Elo11,097次比较
- 原榜Elo
- 1021.16 Elo
- 原榜名次
- 34 名
- 比较次数（非人数）
- 11097 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1011.16 Elo
- Elo95%上界
- 1031.1599999999999 Elo
- 原榜名次区间下界
- 26 名
- 原榜名次区间上界
- 40 名
- 参考价格
- 70 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#35Krea 2 Medium Turbo1019.72 Elo6,060次比较
- 原榜Elo
- 1019.72 Elo
- 原榜名次
- 35 名
- 比较次数（非人数）
- 6060 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1008.72 Elo
- Elo95%上界
- 1030.72 Elo
- 原榜名次区间下界
- 26 名
- 原榜名次区间上界
- 41 名
- 参考价格
- 15 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#36Recraft V4.1 Utility Pro1017.42 Elo15,414次比较
- 原榜Elo
- 1017.42 Elo
- 原榜名次
- 36 名
- 比较次数（非人数）
- 15414 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1008.42 Elo
- Elo95%上界
- 1026.42 Elo
- 原榜名次区间下界
- 30 名
- 原榜名次区间上界
- 42 名
- 参考价格
- 210 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#37HiDream-O1-Image-1.51017.13 Elo12,168次比较
- 原榜Elo
- 1017.13 Elo
- 原榜名次
- 37 名
- 比较次数（非人数）
- 12168 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1006.13 Elo
- Elo95%上界
- 1028.13 Elo
- 原榜名次区间下界
- 29 名
- 原榜名次区间上界
- 42 名
- 参考价格
- 80 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：HiDream-O1-Image-1.5 is the latest version of HiDream O1. Earlier variants, including HiDream-O1-Image, HiDream-O1-Image-Dev, and HiDream-O1-Image-Dev-2604, are benchmarked and listed separately.
模型详情 →#38Seedream 5.0 Lite1016.86 Elo5,605次比较
- 原榜Elo
- 1016.86 Elo
- 原榜名次
- 38 名
- 比较次数（非人数）
- 5605 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1004.86 Elo
- Elo95%上界
- 1028.8600000000001 Elo
- 原榜名次区间下界
- 28 名
- 原榜名次区间上界
- 42 名
- 参考价格
- 35 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#39Ideogram 4.5 (High)1015.07 Elo2,888次比较
- 原榜Elo
- 1015.07 Elo
- 原榜名次
- 39 名
- 比较次数（非人数）
- 2888 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1000.07 Elo
- Elo95%上界
- 1030.0700000000002 Elo
- 原榜名次区间下界
- 26 名
- 原榜名次区间上界
- 44 名
- 参考价格
- 100 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#40Wan2.6 Text to Image1014.20 Elo5,561次比较
- 原榜Elo
- 1014.2 Elo
- 原榜名次
- 40 名
- 比较次数（非人数）
- 5561 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1002.2 Elo
- Elo95%上界
- 1026.2 Elo
- 原榜名次区间下界
- 30 名
- 原榜名次区间上界
- 43 名
- 参考价格
- 30 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Wan 2.6 Text to Image is a first party fine-tune of Wan 2.6 Image specifically for Text to Image generation
模型详情 →#41GPT Image 1 (high)1012.67 Elo5,518次比较
- 原榜Elo
- 1012.67 Elo
- 原榜名次
- 41 名
- 比较次数（非人数）
- 5518 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1000.67 Elo
- Elo95%上界
- 1024.67 Elo
- 原榜名次区间下界
- 31 名
- 原榜名次区间上界
- 44 名
- 参考价格
- 167 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Generated using the OpenAI Images API with `model_id: "gpt-image-1"` and `quality: "high"`. Prompts were pre-pended with "Create the following as a square image:" for Text to Image, and "Edit the following to create a square image:" for Image Editing. We found without pre-pending these instructions, images would not be produced natively as squares.
模型详情 →#42Krea 2 Medium1012.01 Elo14,449次比较
- 原榜Elo
- 1012.01 Elo
- 原榜名次
- 42 名
- 比较次数（非人数）
- 14449 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 1003.01 Elo
- Elo95%上界
- 1021.01 Elo
- 原榜名次区间下界
- 35 名
- 原榜名次区间上界
- 43 名
- 参考价格
- 30 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#43FLUX.2 [pro]1002.80 Elo10,826次比较
- 原榜Elo
- 1002.8 Elo
- 原榜名次
- 43 名
- 比较次数（非人数）
- 10826 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 992.8 Elo
- Elo95%上界
- 1012.8 Elo
- 原榜名次区间下界
- 41 名
- 原榜名次区间上界
- 49 名
- 参考价格
- 30 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#44Ideogram 4.01001.59 Elo10,102次比较
- 原榜Elo
- 1001.59 Elo
- 原榜名次
- 44 名
- 比较次数（非人数）
- 10102 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 991.59 Elo
- Elo95%上界
- 1011.59 Elo
- 原榜名次区间下界
- 41 名
- 原榜名次区间上界
- 50 名
- 参考价格
- 60 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#45FLUX.2 [dev]1000.00 Elo18,040次比较
- 原榜Elo
- 1000 Elo
- 原榜名次
- 45 名
- 比较次数（非人数）
- 18040 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 1000 Elo
- Elo95%上界
- 1000 Elo
- 原榜名次区间下界
- 43 名
- 原榜名次区间上界
- 50 名
- 参考价格
- 12 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#46Ming-Image-0.1-Design999.01 Elo20,241次比较
- 原榜Elo
- 999.01 Elo
- 原榜名次
- 46 名
- 比较次数（非人数）
- 20241 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 990.01 Elo
- Elo95%上界
- 1008.01 Elo
- 原榜名次区间下界
- 43 名
- 原榜名次区间上界
- 51 名
- 参考价格
- 30 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#47P-Image-Ideogram (High)995.42 Elo3,572次比较
- 原榜Elo
- 995.42 Elo
- 原榜名次
- 47 名
- 比较次数（非人数）
- 3572 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 981.42 Elo
- Elo95%上界
- 1009.42 Elo
- 原榜名次区间下界
- 42 名
- 原榜名次区间上界
- 58 名
- 参考价格
- 15 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#48FLUX.2 [dev] Turbo994.04 Elo6,274次比较
- 原榜Elo
- 994.04 Elo
- 原榜名次
- 48 名
- 比较次数（非人数）
- 6274 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 983.04 Elo
- Elo95%上界
- 1005.04 Elo
- 原榜名次区间下界
- 44 名
- 原榜名次区间上界
- 57 名
- 参考价格
- 8 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：FLUX.2 [dev] Turbo is a distilled LoRA adapter for Black Forest Lab's FLUX.2 [dev] that generates images in 8 inference steps
模型详情 →#49Cosmos3-Super-Text2Image (agentic)992.52 Elo11,549次比较
- 原榜Elo
- 992.52 Elo
- 原榜名次
- 49 名
- 比较次数（非人数）
- 11549 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 982.52 Elo
- Elo95%上界
- 1002.52 Elo
- 原榜名次区间下界
- 45 名
- 原榜名次区间上界
- 57 名
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 本源本次显示“No API available”，报价待核验；未采用隐藏raw价unknown，这不证明发布方没有任何API。
模型详情 →#50Luma UNI 1992.19 Elo5,559次比较
- 原榜Elo
- 992.19 Elo
- 原榜名次
- 50 名
- 比较次数（非人数）
- 5559 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 980.19 Elo
- Elo95%上界
- 1004.19 Elo
- 原榜名次区间下界
- 44 名
- 原榜名次区间上界
- 59 名
- 参考价格
- 40.4 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#51P-Image-Ideogram (Very High)991.61 Elo2,776次比较
- 原榜Elo
- 991.61 Elo
- 原榜名次
- 51 名
- 比较次数（非人数）
- 2776 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 976.61 Elo
- Elo95%上界
- 1006.61 Elo
- 原榜名次区间下界
- 43 名
- 原榜名次区间上界
- 61 名
- 参考价格
- 33 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#52Qwen Image 2.0 (2026-03-03)989.54 Elo2,633次比较
- 原榜Elo
- 989.54 Elo
- 原榜名次
- 52 名
- 比较次数（非人数）
- 2633 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 973.54 Elo
- Elo95%上界
- 1005.54 Elo
- 原榜名次区间下界
- 43 名
- 原榜名次区间上界
- 62 名
- 参考价格
- 35 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#53HunyuanImage 3.0 (Tencent Cloud)988.60 Elo2,531次比较
- 原榜Elo
- 988.6 Elo
- 原榜名次
- 53 名
- 比较次数（非人数）
- 2531 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 972.6 Elo
- Elo95%上界
- 1004.6 Elo
- 原榜名次区间下界
- 44 名
- 原榜名次区间上界
- 63 名
- 参考价格
- 32 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Served through Tencent Cloud's hy-image-v3 API with Tencent's prompt revision enabled.
模型详情 →#54FLUX.2 [dev] Flash987.76 Elo5,344次比较
- 原榜Elo
- 987.76 Elo
- 原榜名次
- 54 名
- 比较次数（非人数）
- 5344 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 975.76 Elo
- Elo95%上界
- 999.76 Elo
- 原榜名次区间下界
- 47 名
- 原榜名次区间上界
- 61 名
- 参考价格
- 5 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：FLUX.2 [dev] Flash is the same weights as FLUX.2 [dev] Turbo (a distilled LoRA adapter for Black Forest Lab's FLUX.2 [dev]) that generates images in 5 inference steps
模型详情 →#55HunyuanImage 3.0 Instruct987.36 Elo5,683次比较
- 原榜Elo
- 987.36 Elo
- 原榜名次
- 55 名
- 比较次数（非人数）
- 5683 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 975.36 Elo
- Elo95%上界
- 999.36 Elo
- 原榜名次区间下界
- 47 名
- 原榜名次区间上界
- 61 名
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 本源本次显示“No API available”，报价待核验；未采用隐藏raw价unknown，这不证明发布方没有任何API。 来源说明：Open weights (tencent/HunyuanImage-3.0-Instruct) run on Artificial Analysis hardware with the model's think-and-rewrite prompt step enabled. Tencent's hosted version is listed separately as HunyuanImage 3.0 (Tencent Cloud).
模型详情 →#56Wan 2.7985.56 Elo5,491次比较
- 原榜Elo
- 985.56 Elo
- 原榜名次
- 56 名
- 比较次数（非人数）
- 5491 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 973.56 Elo
- Elo95%上界
- 997.56 Elo
- 原榜名次区间下界
- 48 名
- 原榜名次区间上界
- 62 名
- 参考价格
- 26 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#57Cosmos3-Super-Text2Image985.16 Elo10,887次比较
- 原榜Elo
- 985.16 Elo
- 原榜名次
- 57 名
- 比较次数（非人数）
- 10887 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 975.16 Elo
- Elo95%上界
- 995.16 Elo
- 原榜名次区间下界
- 50 名
- 原榜名次区间上界
- 61 名
- 参考价格
- 40 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#58Nano Banana (Gemini 2.5 Flash Image)985.16 Elo5,312次比较
- 原榜Elo
- 985.16 Elo
- 原榜名次
- 58 名
- 比较次数（非人数）
- 5312 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 973.16 Elo
- Elo95%上界
- 997.16 Elo
- 原榜名次区间下界
- 48 名
- 原榜名次区间上界
- 63 名
- 参考价格
- 39 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#59Ideogram 4.0 Fast (Quality)982.01 Elo5,946次比较
- 原榜Elo
- 982.01 Elo
- 原榜名次
- 59 名
- 比较次数（非人数）
- 5946 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 970.01 Elo
- Elo95%上界
- 994.01 Elo
- 原榜名次区间下界
- 51 名
- 原榜名次区间上界
- 63 名
- 参考价格
- 17.5 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Generated with fal.ai's optimized version of the Ideogram 4.0 open weights model. fal's Fast variant is an FP4-quantized checkpoint trained with quantization-aware distillation, with classifier-free guidance distilled into a single branch for faster inference. Run with fal's QUALITY rendering speed setting.
模型详情 →#60Ideogram 4.0 Instant980.82 Elo5,492次比较
- 原榜Elo
- 980.82 Elo
- 原榜名次
- 60 名
- 比较次数（非人数）
- 5492 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 968.82 Elo
- Elo95%上界
- 992.82 Elo
- 原榜名次区间下界
- 51 名
- 原榜名次区间上界
- 64 名
- 参考价格
- 7.5 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Generated with fal.ai's optimized version of the Ideogram 4.0 open weights model. fal's Instant variant is an FP4-quantized checkpoint trained with quantization-aware distillation, with classifier-free guidance distilled into a single branch and additional timestep distillation to reduce denoising steps.
模型详情 →#61Wan 2.7 Pro979.20 Elo5,498次比较
- 原榜Elo
- 979.2 Elo
- 原榜名次
- 61 名
- 比较次数（非人数）
- 5498 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 967.2 Elo
- Elo95%上界
- 991.2 Elo
- 原榜名次区间下界
- 52 名
- 原榜名次区间上界
- 65 名
- 参考价格
- 75 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#62HiDream-O1-Image975.04 Elo10,662次比较
- 原榜Elo
- 975.04 Elo
- 原榜名次
- 62 名
- 比较次数（非人数）
- 10662 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 965.04 Elo
- Elo95%上界
- 985.04 Elo
- 原榜名次区间下界
- 57 名
- 原榜名次区间上界
- 65 名
- 参考价格
- 40 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：HiDream-O1-Image is the original HiDream O1 model. HiDream has since introduced HiDream-O1-Image-Dev-2604, a fine-tuned version built specifically for Text to Image, as well as HiDream-O1-Image-1.5, which have been benchmarked and listed separately.
模型详情 →#63P-Image-Ideogram (Medium)972.35 Elo3,761次比较
- 原榜Elo
- 972.35 Elo
- 原榜名次
- 63 名
- 比较次数（非人数）
- 3761 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 958.35 Elo
- Elo95%上界
- 986.35 Elo
- 原榜名次区间下界
- 56 名
- 原榜名次区间上界
- 67 名
- 参考价格
- 10 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#64Cosmos3-Super-Text2Image-4Step970.13 Elo9,044次比较
- 原榜Elo
- 970.13 Elo
- 原榜名次
- 64 名
- 比较次数（非人数）
- 9044 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 960.13 Elo
- Elo95%上界
- 980.13 Elo
- 原榜名次区间下界
- 60 名
- 原榜名次区间上界
- 67 名
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 本源本次显示“No API available”，报价待核验；未采用隐藏raw价unknown，这不证明发布方没有任何API。
模型详情 →#65HunyuanImage 3.5 (Preview)969.70 Elo5,586次比较
- 原榜Elo
- 969.7 Elo
- 原榜名次
- 65 名
- 比较次数（非人数）
- 5586 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 957.7 Elo
- Elo95%上界
- 981.7 Elo
- 原榜名次区间下界
- 59 名
- 原榜名次区间上界
- 67 名
- 参考价格
- 24 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Generated through the Tencent Cloud International API (hy-image-v3.5-preview).
模型详情 →#66Ideogram 4.0 Fast963.03 Elo5,438次比较
- 原榜Elo
- 963.03 Elo
- 原榜名次
- 66 名
- 比较次数（非人数）
- 5438 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 951.03 Elo
- Elo95%上界
- 975.03 Elo
- 原榜名次区间下界
- 62 名
- 原榜名次区间上界
- 68 名
- 参考价格
- 10.5 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Generated with fal.ai's optimized version of the Ideogram 4.0 open weights model. fal's Fast variant is an FP4-quantized checkpoint trained with quantization-aware distillation, with classifier-free guidance distilled into a single branch for faster inference. Run with fal's BALANCED rendering speed setting.
模型详情 →#67P-Image-Ideogram (Low)959.20 Elo3,832次比较
- 原榜Elo
- 959.2 Elo
- 原榜名次
- 67 名
- 比较次数（非人数）
- 3832 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 945.2 Elo
- Elo95%上界
- 973.2 Elo
- 原榜名次区间下界
- 63 名
- 原榜名次区间上界
- 69 名
- 参考价格
- 7.5 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#68Qwen Image Plus 2601951.71 Elo2,584次比较
- 原榜Elo
- 951.71 Elo
- 原榜名次
- 68 名
- 比较次数（非人数）
- 2584 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 935.71 Elo
- Elo95%上界
- 967.71 Elo
- 原榜名次区间下界
- 65 名
- 原榜名次区间上界
- 71 名
- 参考价格
- 30 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#69Recraft V4.1 Flash949.34 Elo5,770次比较
- 原榜Elo
- 949.34 Elo
- 原榜名次
- 69 名
- 比较次数（非人数）
- 5770 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 937.34 Elo
- Elo95%上界
- 961.34 Elo
- 原榜名次区间下界
- 67 名
- 原榜名次区间上界
- 71 名
- 参考价格
- 7 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#70GPT Image 1 Mini (medium)935.82 Elo2,333次比较
- 原榜Elo
- 935.82 Elo
- 原榜名次
- 70 名
- 比较次数（非人数）
- 2333 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 918.82 Elo
- Elo95%上界
- 952.82 Elo
- 原榜名次区间下界
- 68 名
- 原榜名次区间上界
- 76 名
- 参考价格
- 11 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：Generated using the OpenAI Images API with `model_id: "gpt-image-1-mini"` and `quality: "medium"`. Prompts were pre-pended with "Create the following as a square image:" for Text to Image, and "Edit the following to create a square image:" for Image Editing. We found without pre-pending these instructions, images would not be produced natively as squares.
模型详情 →#71Z-Image Turbo935.51 Elo8,170次比较
- 原榜Elo
- 935.51 Elo
- 原榜名次
- 71 名
- 比较次数（非人数）
- 8170 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 924.51 Elo
- Elo95%上界
- 946.51 Elo
- 原榜名次区间下界
- 69 名
- 原榜名次区间上界
- 75 名
- 参考价格
- 5 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#72HunyuanImage 3.0931.63 Elo2,443次比较
- 原榜Elo
- 931.63 Elo
- 原榜名次
- 72 名
- 比较次数（非人数）
- 2443 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 915.63 Elo
- Elo95%上界
- 947.63 Elo
- 原榜名次区间下界
- 69 名
- 原榜名次区间上界
- 78 名
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 本源本次显示“No API available”，报价待核验；未采用隐藏raw价unknown，这不证明发布方没有任何API。 来源说明：Open weights (tencent/HunyuanImage-3.0) run on Artificial Analysis hardware with the repository prompt rewrite enabled. Tencent's hosted version is listed separately as HunyuanImage 3.0 (Tencent Cloud).
模型详情 →#73Dreamina 3.1931.11 Elo2,372次比较
- 原榜Elo
- 931.11 Elo
- 原榜名次
- 73 名
- 比较次数（非人数）
- 2372 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 914.11 Elo
- Elo95%上界
- 948.11 Elo
- 原榜名次区间下界
- 69 名
- 原榜名次区间上界
- 78 名
- 参考价格
- 30 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#74FLUX.2 [klein] 9B929.45 Elo10,730次比较
- 原榜Elo
- 929.45 Elo
- 原榜名次
- 74 名
- 比较次数（非人数）
- 10730 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 919.45 Elo
- Elo95%上界
- 939.45 Elo
- 原榜名次区间下界
- 70 名
- 原榜名次区间上界
- 76 名
- 参考价格
- 15 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#75FLUX.1 Kontext [max]928.67 Elo5,276次比较
- 原榜Elo
- 928.67 Elo
- 原榜名次
- 75 名
- 比较次数（非人数）
- 5276 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 916.67 Elo
- Elo95%上界
- 940.67 Elo
- 原榜名次区间下界
- 70 名
- 原榜名次区间上界
- 77 名
- 参考价格
- 80 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#76Kling Image 3.0 Omni917.99 Elo2,672次比较
- 原榜Elo
- 917.99 Elo
- 原榜名次
- 76 名
- 比较次数（非人数）
- 2672 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 900.99 Elo
- Elo95%上界
- 934.99 Elo
- 原榜名次区间下界
- 72 名
- 原榜名次区间上界
- 83 名
- 参考价格
- 28 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#77FLUX.1 Kontext [pro]917.44 Elo2,229次比较
- 原榜Elo
- 917.44 Elo
- 原榜名次
- 77 名
- 比较次数（非人数）
- 2229 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 899.44 Elo
- Elo95%上界
- 935.44 Elo
- 原榜名次区间下界
- 72 名
- 原榜名次区间上界
- 83 名
- 参考价格
- 40 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#78ERNIE Image Turbo914.28 Elo10,560次比较
- 原榜Elo
- 914.28 Elo
- 原榜名次
- 78 名
- 比较次数（非人数）
- 10560 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 904.28 Elo
- Elo95%上界
- 924.28 Elo
- 原榜名次区间下界
- 75 名
- 原榜名次区间上界
- 82 名
- 参考价格
- 10.49 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#79Ideogram 3.0910.57 Elo2,333次比较
- 原榜Elo
- 910.57 Elo
- 原榜名次
- 79 名
- 比较次数（非人数）
- 2333 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 892.57 Elo
- Elo95%上界
- 928.57 Elo
- 原榜名次区间下界
- 74 名
- 原榜名次区间上界
- 85 名
- 参考价格
- 60 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#80P-Image-Ideogram (Very Low)908.87 Elo3,309次比较
- 原榜Elo
- 908.87 Elo
- 原榜名次
- 80 名
- 比较次数（非人数）
- 3309 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 893.87 Elo
- Elo95%上界
- 923.87 Elo
- 原榜名次区间下界
- 76 名
- 原榜名次区间上界
- 85 名
- 参考价格
- 3 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#81ERNIE Image907.06 Elo10,760次比较
- 原榜Elo
- 907.06 Elo
- 原榜名次
- 81 名
- 比较次数（非人数）
- 10760 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 897.06 Elo
- Elo95%上界
- 917.06 Elo
- 原榜名次区间下界
- 77 名
- 原榜名次区间上界
- 84 名
- 参考价格
- 31.46 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#82Qwen Image903.41 Elo2,339次比较
- 原榜Elo
- 903.41 Elo
- 原榜名次
- 82 名
- 比较次数（非人数）
- 2339 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 885.41 Elo
- Elo95%上界
- 921.41 Elo
- 原榜名次区间下界
- 76 名
- 原榜名次区间上界
- 87 名
- 参考价格
- 20 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#83P-Image899.57 Elo2,351次比较
- 原榜Elo
- 899.57 Elo
- 原榜名次
- 83 名
- 比较次数（非人数）
- 2351 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 881.57 Elo
- Elo95%上界
- 917.57 Elo
- 原榜名次区间下界
- 77 名
- 原榜名次区间上界
- 87 名
- 参考价格
- 5 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#84FLUX.1 Krea [dev]896.25 Elo515次比较
- 原榜Elo
- 896.25 Elo
- 原榜名次
- 84 名
- 比较次数（非人数）
- 515 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 857.25 Elo
- Elo95%上界
- 935.25 Elo
- 原榜名次区间下界
- 72 名
- 原榜名次区间上界
- 91 名
- 参考价格
- 25 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#85Vidu Q2895.82 Elo2,524次比较
- 原榜Elo
- 895.82 Elo
- 原榜名次
- 85 名
- 比较次数（非人数）
- 2524 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 878.82 Elo
- Elo95%上界
- 912.82 Elo
- 原榜名次区间下界
- 79 名
- 原榜名次区间上界
- 88 名
- 参考价格
- 30 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#86Z-Image Base891.33 Elo2,500次比较
- 原榜Elo
- 891.33 Elo
- 原榜名次
- 86 名
- 比较次数（非人数）
- 2500 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 874.33 Elo
- Elo95%上界
- 908.33 Elo
- 原榜名次区间下界
- 80 名
- 原榜名次区间上界
- 88 名
- 参考价格
- 10 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#87GLM-Image885.40 Elo2,457次比较
- 原榜Elo
- 885.4 Elo
- 原榜名次
- 87 名
- 比较次数（非人数）
- 2457 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 867.4 Elo
- Elo95%上界
- 903.4 Elo
- 原榜名次区间下界
- 82 名
- 原榜名次区间上界
- 89 名
- 参考价格
- 50 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#88Lucid Origin Ultra877.35 Elo2,689次比较
- 原榜Elo
- 877.35 Elo
- 原榜名次
- 88 名
- 比较次数（非人数）
- 2689 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 860.35 Elo
- Elo95%上界
- 894.35 Elo
- 原榜名次区间下界
- 85 名
- 原榜名次区间上界
- 90 名
- 参考价格
- 86.71 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#89FLUX1.1 [pro]870.19 Elo2,247次比较
- 原榜Elo
- 870.19 Elo
- 原榜名次
- 89 名
- 比较次数（非人数）
- 2247 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 851.19 Elo
- Elo95%上界
- 889.19 Elo
- 原榜名次区间下界
- 86 名
- 原榜名次区间上界
- 92 名
- 参考价格
- 40 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#90Lucid Origin Fast860.01 Elo2,611次比较
- 原榜Elo
- 860.01 Elo
- 原榜名次
- 90 名
- 比较次数（非人数）
- 2611 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 842.01 Elo
- Elo95%上界
- 878.01 Elo
- 原榜名次区间下界
- 88 名
- 原榜名次区间上界
- 94 名
- 参考价格
- 17.94 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#91LongCat Image856.05 Elo2,146次比较
- 原榜Elo
- 856.05 Elo
- 原榜名次
- 91 名
- 比较次数（非人数）
- 2146 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 837.05 Elo
- Elo95%上界
- 875.05 Elo
- 原榜名次区间下界
- 88 名
- 原榜名次区间上界
- 95 名
- 参考价格
- 130 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#92FLUX1.1 [pro] Ultra854.73 Elo2,284次比较
- 原榜Elo
- 854.73 Elo
- 原榜名次
- 92 名
- 比较次数（非人数）
- 2284 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 835.73 Elo
- Elo95%上界
- 873.73 Elo
- 原榜名次区间下界
- 89 名
- 原榜名次区间上界
- 96 名
- 参考价格
- 60 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#93FIBO848.86 Elo2,258次比较
- 原榜Elo
- 848.86 Elo
- 原榜名次
- 93 名
- 比较次数（非人数）
- 2258 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 829.86 Elo
- Elo95%上界
- 867.86 Elo
- 原榜名次区间下界
- 89 名
- 原榜名次区间上界
- 97 名
- 参考价格
- 40 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#94Luma Photon843.26 Elo2,251次比较
- 原榜Elo
- 843.26 Elo
- 原榜名次
- 94 名
- 比较次数（非人数）
- 2251 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 824.26 Elo
- Elo95%上界
- 862.26 Elo
- 原榜名次区间下界
- 90 名
- 原榜名次区间上界
- 98 名
- 参考价格
- 19 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#95FLUX.2 [klein] 4B837.21 Elo6,418次比较
- 原榜Elo
- 837.21 Elo
- 原榜名次
- 95 名
- 比较次数（非人数）
- 6418 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 825.21 Elo
- Elo95%上界
- 849.21 Elo
- 原榜名次区间下界
- 93 名
- 原榜名次区间上界
- 98 名
- 参考价格
- 14 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#96Phoenix 0.9 Ultra833.38 Elo2,214次比较
- 原榜Elo
- 833.38 Elo
- 原榜名次
- 96 名
- 比较次数（非人数）
- 2214 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 813.38 Elo
- Elo95%上界
- 853.38 Elo
- 原榜名次区间下界
- 92 名
- 原榜名次区间上界
- 100 名
- 参考价格
- 76.24 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#97Phoenix 1.0 Fast826.70 Elo2,151次比较
- 原榜Elo
- 826.7 Elo
- 原榜名次
- 97 名
- 比较次数（非人数）
- 2151 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 806.7 Elo
- Elo95%上界
- 846.7 Elo
- 原榜名次区间下界
- 93 名
- 原榜名次区间上界
- 101 名
- 参考价格
- 19.43 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#98SRPO825.72 Elo2,154次比较
- 原榜Elo
- 825.72 Elo
- 原榜名次
- 98 名
- 比较次数（非人数）
- 2154 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 805.72 Elo
- Elo95%上界
- 845.72 Elo
- 原榜名次区间下界
- 94 名
- 原榜名次区间上界
- 101 名
- 参考价格
- 26 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#99FLUX.1 [dev]822.60 Elo511次比较
- 原榜Elo
- 822.6 Elo
- 原榜名次
- 99 名
- 比较次数（非人数）
- 511 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 777.6 Elo
- Elo95%上界
- 867.6 Elo
- 原榜名次区间下界
- 90 名
- 原榜名次区间上界
- 101 名
- 参考价格
- 25 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#100FIBO Lite813.06 Elo3,300次比较
- 原榜Elo
- 813.06 Elo
- 原榜名次
- 100 名
- 比较次数（非人数）
- 3300 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 797.06 Elo
- Elo95%上界
- 829.06 Elo
- 原榜名次区间下界
- 97 名
- 原榜名次区间上界
- 101 名
- 参考价格
- 20 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#101HiDream-O1-Image-Dev809.87 Elo2,323次比较
- 原榜Elo
- 809.87 Elo
- 原榜名次
- 101 名
- 比较次数（非人数）
- 2323 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 789.87 Elo
- Elo95%上界
- 829.87 Elo
- 原榜名次区间下界
- 97 名
- 原榜名次区间上界
- 101 名
- 参考价格
- 5 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators. 来源说明：HiDream-O1-Image-Dev is the faster, distilled version of HiDream O1. HiDream has since introduced HiDream-O1-Image-Dev-2604, a fine-tuned version built specifically for Text to Image, as well as HiDream-O1-Image-1.5, which have been benchmarked and listed separately.
模型详情 →#102Luma Photon Flash768.83 Elo2,052次比较
- 原榜Elo
- 768.83 Elo
- 原榜名次
- 102 名
- 比较次数（非人数）
- 2052 次比较
- 开放权重资料
- 0 标记
- Elo95%下界
- 745.83 Elo
- Elo95%上界
- 791.83 Elo
- 原榜名次区间下界
- 102 名
- 原榜名次区间上界
- 103 名
- 参考价格
- 5 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#103FLUX.1 [schnell]739.50 Elo3,045次比较
- 原榜Elo
- 739.5 Elo
- 原榜名次
- 103 名
- 比较次数（非人数）
- 3045 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 719.5 Elo
- Elo95%上界
- 759.5 Elo
- 原榜名次区间下界
- 102 名
- 原榜名次区间上界
- 104 名
- 参考价格
- 3 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#104OmniGen V2717.44 Elo2,043次比较
- 原榜Elo
- 717.44 Elo
- 原榜名次
- 104 名
- 比较次数（非人数）
- 2043 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 691.44 Elo
- Elo95%上界
- 743.44 Elo
- 原榜名次区间下界
- 103 名
- 原榜名次区间上界
- 104 名
- 参考价格
- 150 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →#105Bagel593.57 Elo2,081次比较
- 原榜Elo
- 593.57 Elo
- 原榜名次
- 105 名
- 比较次数（非人数）
- 2081 次比较
- 开放权重资料
- 1 标记
- Elo95%下界
- 558.57 Elo
- Elo95%上界
- 628.57 Elo
- 原榜名次区间下界
- 105 名
- 原榜名次区间上界
- 105 名
- 参考价格
- 100 $/千张
AA-Image-T2I v2.1 Overall；text-to-image；not-applicable；Artificial Analysis原榜偏好评测，Elo仅在同任务/协议/版本比较。比较次数非人数；原评测截止日及原分数更新时间未提供。价格是第三方参考，不是独立核实的官方报价。开放权重不证明开源/商业许可。来源价格条件原文：API Pricing reflects the cost to generate 1,000 images on the model creator's API at 1024x1024 with the model's default settings。来源评审说明原文（What are the top models on AA-Image-T2I v2.1?）：The top models on AA-Image-T2I v2.1 by Elo rating are: 1. GPT Image 2.5 Sunburst (max) (Elo 1222), 2. GPT Image 2.5 Flare (max) (Elo 1207), 3. GPT Image 2 (high) (Elo 1186), 4. Nano Banana 2.1 (Elo 1176), and 5. MAI-Image-2.6 (Elo 1166). Rankings are based on blind votes from our validated private evaluators.
模型详情 →