GraphicDesignBench
Models analyze and produce graphic designs in 50 tasks spanning layout, typography, infographics, design semantics, and animation. The suite uses professional layered design templates.
The April 2026 paper reports task-specific metrics. Veo 2.0 led Sora on motion trajectory generation at 60% motion accuracy.
- Made by
- LICA World
- Released
- April 2026
- Kind
- General visual questions
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Capability
- Subject
- Arts
- Language
- English
- Interaction
- Single turn
- Items from
- Real work and use
- Refresh
- One fixed set
- Run by
- Anyone
- Occupations
- Arts, design, entertainment, sports and media
- In views
- Vision, Image, video and 3D generation
- Strongest model tested
- GPT-5.4 #28 on the capability index
- Last model added
- March 2026
- Scores refreshed
- November 20th, 2026
GPT-5.4 leads with 0.94 among models released by March 2026
Best score per model from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. The paper reports task-specific measures. The headline is aspect-ratio classification accuracy on 989 designs, the first measure in Table 5. Where a model ran with several settings or harnesses, we keep its best.
The newest model with a score here came out in March 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
- 1 GPT-5.4 0.94
- 2
Gemini 3.1 Pro 0.24
- 3
Gemini 3.1 Flash-Lite 0.24
- 4 Claude Opus 4.6 0.09
Source: The GraphicDesignBench paper
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5.4 | 0.94 | Mar 2026 | Official leaderboard | |
| 2 | | 0.24 | Feb 2026 | Official leaderboard | |
| 3 | | 0.24 | Mar 2026 | Official leaderboard | |
| 4 | Claude Opus 4.6 | 0.09 | Feb 2026 | Official leaderboard |
Rankings by part
We read 12 parts from the benchmark’s paper, each ranking models on one part of the test, such as Aspect ratio accuracy, Aspect ratio F1 and Element counting MAE. GPT-5.4 leads 9 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Aspect ratio accuracy
4 models we track have a score here.
- 1 GPT-5.4 0.94
- 2
Gemini 3.1 Pro 0.24
- 3
Gemini 3.1 Flash-Lite 0.24
- 4 Claude Opus 4.6 0.09
Aspect ratio F1
4 models we track have a score here.
- 1 GPT-5.4 0.68
- 2 Claude Opus 4.6 ↑2 0.18
- 3
Gemini 3.1 Flash-Lite 0.10
- 4
Gemini 3.1 Pro ↓2 0.09
Element counting MAE
4 models we track have a score here.
- 1 GPT-5.4 5.81
- 2 Claude Opus 4.6 ↑2 6.46
- 3
Gemini 3.1 Pro ↓1 12.3
- 4
Gemini 3.1 Flash-Lite ↓1 12.5
Element counting MSE
4 models we track have a score here.
- 1 GPT-5.4 134.0
- 2 Claude Opus 4.6 ↑2 150.7
- 3
Gemini 3.1 Pro ↓1 342.9
- 4
Gemini 3.1 Flash-Lite ↓1 375.7
Component type accuracy
4 models we track have a score here.
- 1 GPT-5.4 0.46
- 2
Gemini 3.1 Pro 0.28
- 3
Gemini 3.1 Flash-Lite 0.25
- 4 Claude Opus 4.6 0.07
Component type F1
4 models we track have a score here.
- 1 GPT-5.4 0.36
- 2
Gemini 3.1 Pro 0.24
- 3
Gemini 3.1 Flash-Lite 0.22
- 4 Claude Opus 4.6 0.10
Component detection mAP at 0.5
4 models we track have a score here.
- 1 GPT-5.4 0.06
- 2 Claude Opus 4.6 ↑2 0.02
- 3
Gemini 3.1 Pro ↓1 0.01
- 4
Gemini 3.1 Flash-Lite ↓1 0.01
Component detection mAP at 0.5 to 0.95
4 models we track have a score here.
- 1 GPT-5.4 0.04
- 2 Claude Opus 4.6 ↑2 0.01
- 3
Gemini 3.1 Pro ↓1 0.00
- 4
Gemini 3.1 Flash-Lite ↓1 0.00
Text component AP
4 models we track have a score here.
- 1 GPT-5.4 0.07
- 2 Claude Opus 4.6 ↑2 0.02
- 3
Gemini 3.1 Flash-Lite 0.00
- 4
Gemini 3.1 Pro ↓2 0.00
Image component AP
4 models we track have a score here.
- 1 Claude Opus 4.6 ↑3 0.03
- 2 GPT-5.4 ↓1 0.02
- 3
Gemini 3.1 Flash-Lite 0.01
- 4
Gemini 3.1 Pro ↓2 0.00
Layer order Kendall tau
4 models we track have a score here.
- 1
Gemini 3.1 Flash-Lite ↑2 0.57
- 2 Claude Opus 4.6 ↑2 0.54
- 3
Gemini 3.1 Pro ↓1 0.49
- 4 GPT-5.4 ↓3 0.49
Layer order Spearman rho
4 models we track have a score here.
- 1 Claude Opus 4.6 ↑3 0.57
- 2
Gemini 3.1 Flash-Lite ↑1 0.57
- 3 GPT-5.4 ↓2 0.54
- 4
Gemini 3.1 Pro ↓2 0.52
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The GraphicDesignBench paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official GraphicDesignBench 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 8 months ago, in March 2026, for GPT-5.4. 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 benchmark’s paper (4 entries).
- Note Models scored Only 4 models we track.
- Warning Newest score The newest came 8 months ago, in March 2026, for GPT-5.4.
- Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.4 (#28).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 4 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.