GenPuzzle
Models generate images that solve visual puzzles such as mazes, tangrams, and Sudoku. The authors select an automated judge using human reference scores.
- Made by
- Changpeng Zhao and colleagues
- Released
- September 2026
- Kind
- Image generation and editing
- Level
- School
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Vision, Image, video and 3D generation
- Last model added
- February 2026
- Scores refreshed
- November 20th, 2026
Seedream 5.0 Pro leads with 40.6 among models released by February 2026
Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best.
The newest model with a score here came out in February 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
Source: Official GenPuzzle leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 40.6 | Gemini 3.1 Pro judge | Official leaderboard | |
| 2 | | 32.4 | Gemini 3.1 Pro judge | Feb 2026 | Official leaderboard |
| 3 | GPT-Image-2 | 29.4 | Gemini 3.1 Pro judge | Official leaderboard |
Rankings by part
We read 12 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Figure Completion, Spatial Generation and Maze: Beginner. Seedream 5.0 Pro leads 9 of them, and Gemini 3.1 Flash Image leads the other 3. Arrows show how far a model moved from its overall rank among the models we track.
Figure Completion
3 models we track have a score here.
- 1
Seedream 5.0 Pro 30.9
- 2
Gemini 3.1 Flash Image 19.8
- 3 GPT-Image-2 15.1
Spatial Generation
3 models we track have a score here.
- 1
Gemini 3.1 Flash Image ↑1 17.3
- 2 GPT-Image-2 ↑1 13.7
- 3
Seedream 5.0 Pro ↓2 2.07
Maze: Beginner
3 models we track have a score here.
- 1
Seedream 5.0 Pro 33.3
- 2 GPT-Image-2 ↑1 29.7
- 3
Gemini 3.1 Flash Image ↓1 29.7
Maze: Intermediate
3 models we track have a score here.
- 1
Seedream 5.0 Pro 31.8
- 2
Gemini 3.1 Flash Image 31.8
- 3 GPT-Image-2 30.2
Maze: Advanced
3 models we track have a score here.
- 1
Gemini 3.1 Flash Image ↑1 29.2
- 2 GPT-Image-2 ↑1 28.6
- 3
Seedream 5.0 Pro ↓2 28.6
Sudoku
3 models we track have a score here.
- 1
Seedream 5.0 Pro 96.2
- 2
Gemini 3.1 Flash Image 72.2
- 3 GPT-Image-2 60.3
Nonogram
3 models we track have a score here.
- 1
Seedream 5.0 Pro 49.8
- 2
Gemini 3.1 Flash Image 36.9
- 3 GPT-Image-2 29.1
Tangram
3 models we track have a score here.
- 1
Seedream 5.0 Pro 45.8
- 2 GPT-Image-2 ↑1 17.6
- 3
Gemini 3.1 Flash Image ↓1 16.2
Board Games
3 models we track have a score here.
- 1
Seedream 5.0 Pro 66.7
- 2
Gemini 3.1 Flash Image 44.6
- 3 GPT-Image-2 42.0
Matchsticks
3 models we track have a score here.
- 1
Seedream 5.0 Pro 10.7
- 2
Gemini 3.1 Flash Image 3.00
- 3 GPT-Image-2 0.33
Orthographic
3 models we track have a score here.
- 1
Seedream 5.0 Pro 53.1
- 2 GPT-Image-2 ↑1 50.5
- 3
Gemini 3.1 Flash Image ↓1 49.0
Math Visual Proof
3 models we track have a score here.
- 1
Gemini 3.1 Flash Image ↑1 38.7
- 2
Seedream 5.0 Pro ↓1 38.0
- 3 GPT-Image-2 35.7
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official GenPuzzle leaderboard
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 Newest score check. The newest came 9 months ago, in February 2026, for Gemini 3.1 Flash Image. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 1 pass 3 notes
- Pass Independent runs Scores from the official leaderboard (3 entries).
- Note Models scored Only 3 models we track.
- Warning Newest score The newest came 9 months ago, in February 2026, for Gemini 3.1 Flash Image.
- 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 huggingface.co: 3 rows for models we track at the last read.
- Pass Last read November 20th, 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.