SuperCLUE-Edit
Models edit an input image according to a Chinese instruction. The evaluation measures whether the requested change preserves relevant details.
SuperCLUE publishes scores for 13 models in its 2026 results workbook.
- Made by
- SuperCLUE
- Released
- October 2025
- Kind
- Image generation and editing
- Level
- Everyday
- Scored
- Graded by a model
- Answers
- Open-ended
- Measures
- Capability
- Language
- Chinese
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- A host
- In views
- Vision, Image, video and 3D generation
- Scores refreshed
- November 21st, 2026
ArtArk tops the leaderboard with 83.6 GPT-Image-2.5 Sunburst leads with 82.5
Best score per entry from the benchmark's own leaderboard, read on November 21st, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is GPT-Image-2.5 Sunburst, at 82.5. This view shows only models we track, each at its best.
- 83.6 ArtArk
- 82.5 GPT-Image-2.5 Sunburst
- 82.3 Nano Banana 2.1
- 81.7 GPT-Image-2.5 Flare
- 81.1 Grok Imagine Image 2.0
- 80.8 Doubao-Seedream-5.0-pro
- 80.5
Qwen-Image-3.0-Pro
- 80.3 GPT-Image-2
- 79.8
MAI-Image-2.6
- 79.7 Nano Banana Pro
- 77.8
Wan 2.7 Image Pro
- 74.6 Hunyuan Image 3.5 Preview
- 70.5 Kling Image 3.0 Omni
- 1 ArtArk 83.6
- 2 GPT-Image-2.5 Sunburst 82.5
- 3 Nano Banana 2.1 82.3
- 4 GPT-Image-2.5 Flare 81.7
- 5 Grok Imagine Image 2.0 81.1
- 6 Doubao-Seedream-5.0-pro 80.8
- 7
Qwen-Image-3.0-Pro 80.5
- 8 GPT-Image-2 80.3
- 9
MAI-Image-2.6 79.8
- 10 Nano Banana Pro 79.7
- 11
Wan 2.7 Image Pro 77.8
- 12 Hunyuan Image 3.5 Preview 74.6
- 13 Kling Image 3.0 Omni 70.5
Source: Official SuperCLUE-Edit leaderboard
- 1 GPT-Image-2.5 Sunburst 82.5
- 2 GPT-Image-2.5 Flare 81.7
- 3 Grok Imagine Image 2.0 81.1
- 4
Qwen-Image-3.0-Pro 80.5
- 5 GPT-Image-2 80.3
- 6
MAI-Image-2.6 79.8
- 7
Wan 2.7 Image Pro 77.8
Source: Official SuperCLUE-Edit leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | ArtArk closed weights | 83.6 | Official leaderboard | ||
| 2 | GPT-Image-2.5 Sunburst | 82.5 | Official leaderboard | ||
| 3 | Nano Banana 2.1 closed weights | 82.3 | Official leaderboard | ||
| 4 | GPT-Image-2.5 Flare | 81.7 | Official leaderboard | ||
| 5 | Grok Imagine Image 2.0 | 81.1 | Official leaderboard | ||
| 6 | Doubao-Seedream-5.0-pro closed weights | 80.8 | Official leaderboard | ||
| 7 | | 80.5 | Official leaderboard | ||
| 8 | GPT-Image-2 | 80.3 | Official leaderboard | ||
| 9 | | 79.8 | Official leaderboard | ||
| 10 | Nano Banana Pro closed weights | 79.7 | Official leaderboard | ||
| 11 | | 77.8 | Official leaderboard | ||
| 12 | Hunyuan Image 3.5 Preview closed weights | 74.6 | Official leaderboard | ||
| 13 | Kling Image 3.0 Omni closed weights | 70.5 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-Image-2.5 Sunburst | 82.5 | Official leaderboard | ||
| 2 | GPT-Image-2.5 Flare | 81.7 | Official leaderboard | ||
| 3 | Grok Imagine Image 2.0 | 81.1 | Official leaderboard | ||
| 4 | | 80.5 | Official leaderboard | ||
| 5 | GPT-Image-2 | 80.3 | Official leaderboard | ||
| 6 | | 79.8 | Official leaderboard | ||
| 7 | | 77.8 | Official leaderboard |
Rankings by part
We read 11 parts from the benchmark’s own leaderboard in 2 groups: 通用能力 and 场景应用. Each ranks models on one part of the test. GPT-Image-2.5 Sunburst leads 6 of them, and 4 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
通用能力
7 models we track have a score here.
- 1 GPT-Image-2 ↑4 82.8
- 2 GPT-Image-2.5 Flare 82.6
- 3
Wan 2.7 Image Pro ↑4 82.3
- 4 Grok Imagine Image 2.0 ↓1 82.0
- 5 GPT-Image-2.5 Sunburst ↓4 81.9
- 6
Qwen-Image-3.0-Pro ↓2 81.4
- 7
MAI-Image-2.6 ↓1 81.2
场景应用
7 models we track have a score here.
- 1 GPT-Image-2.5 Sunburst 83.0
- 2 GPT-Image-2.5 Flare 80.8
- 3 Grok Imagine Image 2.0 80.2
- 4
Qwen-Image-3.0-Pro 79.5
- 5
MAI-Image-2.6 ↑1 78.5
- 6 GPT-Image-2 ↓1 77.7
- 7
Wan 2.7 Image Pro 73.3
分数
7 models we track have a score here.
- 1 GPT-Image-2 ↑4 82.8
- 2 GPT-Image-2.5 Flare 82.6
- 3
Wan 2.7 Image Pro ↑4 82.3
- 4 Grok Imagine Image 2.0 ↓1 82.0
- 5 GPT-Image-2.5 Sunburst ↓4 81.9
- 6
Qwen-Image-3.0-Pro ↓2 81.4
- 7
MAI-Image-2.6 ↓1 81.2
对象编辑
7 models we track have a score here.
- 1 GPT-Image-2.5 Flare ↑1 89.3
- 2 GPT-Image-2.5 Sunburst ↓1 89.1
- 3
MAI-Image-2.6 ↑3 88.9
- 4 GPT-Image-2 ↑1 88.9
- 5
Wan 2.7 Image Pro ↑2 88.4
- 6 Grok Imagine Image 2.0 ↓3 86.2
- 7
Qwen-Image-3.0-Pro ↓3 85.6
文本设计
7 models we track have a score here.
- 1 GPT-Image-2.5 Sunburst 91.7
- 2 GPT-Image-2 ↑3 91.4
- 3
MAI-Image-2.6 ↑3 90.6
- 4 GPT-Image-2.5 Flare ↓2 89.5
- 5
Qwen-Image-3.0-Pro ↓1 88.9
- 6 Grok Imagine Image 2.0 ↓3 88.8
- 7
Wan 2.7 Image Pro 88.4
场景构建
7 models we track have a score here.
- 1
MAI-Image-2.6 ↑5 88.9
- 2 GPT-Image-2 ↑3 88.9
- 3 GPT-Image-2.5 Sunburst ↓2 88.7
- 4 GPT-Image-2.5 Flare ↓2 88.6
- 5 Grok Imagine Image 2.0 ↓2 88.5
- 6
Wan 2.7 Image Pro ↑1 88.4
- 7
Qwen-Image-3.0-Pro ↓3 88.1
图像质量
7 models we track have a score here.
- 1 GPT-Image-2.5 Sunburst 80.1
- 2 Grok Imagine Image 2.0 ↑1 80.0
- 3 GPT-Image-2.5 Flare ↓1 79.0
- 4 GPT-Image-2 ↑1 78.0
- 5
Qwen-Image-3.0-Pro ↓1 77.8
- 6
Wan 2.7 Image Pro ↑1 77.6
- 7
MAI-Image-2.6 ↓1 77.4
复杂视觉推理
7 models we track have a score here.
- 1
Wan 2.7 Image Pro ↑6 68.9
- 2 GPT-Image-2 ↑3 66.7
- 3 GPT-Image-2.5 Flare ↓1 66.7
- 4 Grok Imagine Image 2.0 ↓1 66.7
- 5
Qwen-Image-3.0-Pro ↓1 66.7
- 6 GPT-Image-2.5 Sunburst ↓5 60.0
- 7
MAI-Image-2.6 ↓1 60.0
分数
7 models we track have a score here.
- 1 GPT-Image-2.5 Sunburst 83.0
- 2 GPT-Image-2.5 Flare 80.8
- 3 Grok Imagine Image 2.0 80.2
- 4
Qwen-Image-3.0-Pro 79.5
- 5
MAI-Image-2.6 ↑1 78.5
- 6 GPT-Image-2 ↓1 77.7
- 7
Wan 2.7 Image Pro 73.3
创作能力
7 models we track have a score here.
- 1 GPT-Image-2.5 Sunburst 84.4
- 2 GPT-Image-2 ↑3 84.0
- 3 GPT-Image-2.5 Flare ↓1 84.0
- 4 Grok Imagine Image 2.0 ↓1 83.0
- 5
Wan 2.7 Image Pro ↑2 81.0
- 6
MAI-Image-2.6 80.3
- 7
Qwen-Image-3.0-Pro ↓3 79.4
工程能力
7 models we track have a score here.
- 1 GPT-Image-2.5 Sunburst 81.7
- 2
Qwen-Image-3.0-Pro ↑2 79.5
- 3 GPT-Image-2.5 Flare ↓1 77.6
- 4 Grok Imagine Image 2.0 ↓1 77.4
- 5
MAI-Image-2.6 ↑1 76.8
- 6 GPT-Image-2 ↓1 71.5
- 7
Wan 2.7 Image Pro 65.6
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
No dated values yet.
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Vision
- MMMU In use 11.5K college-level multimodal questions across 30 subjects. Multiple-choice and open accuracy.
Qwen 3.6 Plus 86%
- MMMU-Pro In use Harder MMMU: college-level multimodal questions with more options and vision-only inputs to remove text shortcuts. GPT-5.6 Sol 84.6%
- CharXiv In use Chart understanding on real arXiv figures. Reasoning questions require synthesizing information from complex charts. Claude Fable 5 93.5%
- MathVista In use Math reasoning in visual contexts (charts, geometry, puzzles) from 31 datasets. Answer accuracy.
Seed2.1 Pro 90.7%
- MATH-Vision (MathVision) In use 3,040 competition math problems with visual context. Answer accuracy.
Qwen 3.8 Max 97.7%
- MMStar In use Models answer 1,500 human-curated image questions that cannot be solved without the image, across 6 core capabilities. Scored by accuracy.
Qwen 3.8 Max 85.9%
Data audit
7 of 8 checks pass, 1 warning. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark
Catalog record 1 warning 1 pass
- Warning Task access We don’t know whether the tasks are public.
- Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
- Pass Independent runs Scores from the official leaderboard (13 entries).
- Pass Models scored 7 models we track, plus 6 other entries on its leaderboard.
- Note Newest score No scores yet.
- Note Frontier models None of its models are on the capabilities index.
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on superclueai.com: 7 rows for models we track and 6 other entries at the last read.
- Pass Last read November 21st, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.