MDMR-Bench
Models generate one image from several references. A model judge checks whether the output preserves requested elements, follows instructions, and combines them correctly.
- Made by
- Bowen Zheng and colleagues
- Released
- May 2026
- Kind
- Image generation and editing
- Level
- Everyday
- 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
GPT-Image-2 leads with 88.8% among models released by February 2026 GPT-Image-2 leads with 88.8% among models released by February 2026
Best score per entry from the benchmark's own leaderboard, read on November 20th, 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. This view shows only models we track, each at 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
- 1 GPT-Image-2 88.8%
- 2 Nano Banana Pro 85.1%
- 3
Gemini 3.1 Flash Image 81.2%
- 4 Seedream 5.0 78.8%
- 5
Seedream 4.0 77.8%
- 6 GPT-Image-1.5 77.5%
- 7 FLUX.2 [dev] 51.3%
- 8
Qwen-Image-Edit (2511) 49.9%
- 9
BAGEL 41.9%
- 10 DreamOmni2 41.9%
- 11 OmniGen2 28.6%
Source: Official MDMR-Bench leaderboard
- 1 GPT-Image-2 88.8%
- 2
Gemini 3.1 Flash Image 81.2%
- 3
Seedream 4.0 77.8%
- 4 GPT-Image-1.5 77.5%
- 5 FLUX.2 [dev] 51.3%
- 6
Qwen-Image-Edit (2511) 49.9%
- 7
BAGEL 41.9%
Source: Official MDMR-Bench leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-Image-2 | 88.8% | Official leaderboard | ||
| 2 | Nano Banana Pro closed weights | 85.1% | Official leaderboard | ||
| 3 | | 81.2% | Feb 2026 | Official leaderboard | |
| 4 | Seedream 5.0 closed weights | 78.8% | Official leaderboard | ||
| 5 | | 77.8% | Official leaderboard | ||
| 6 | GPT-Image-1.5 | 77.5% | Official leaderboard | ||
| 7 | FLUX.2 [dev] | 51.3% | Official leaderboard | ||
| 8 | | 49.9% | Official leaderboard | ||
| 9 | | 41.9% | Official leaderboard | ||
| 10 | DreamOmni2 open weights | 41.9% | Official leaderboard | ||
| 11 | OmniGen2 open weights | 28.6% | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-Image-2 | 88.8% | Official leaderboard | ||
| 2 | | 81.2% | Feb 2026 | Official leaderboard | |
| 3 | | 77.8% | Official leaderboard | ||
| 4 | GPT-Image-1.5 | 77.5% | Official leaderboard | ||
| 5 | FLUX.2 [dev] | 51.3% | Official leaderboard | ||
| 6 | | 49.9% | Official leaderboard | ||
| 7 | | 41.9% | Official leaderboard |
Rankings by part
We read 10 parts from the benchmark’s own leaderboard in 2 groups: dimension and task. Each ranks models on one part of the test. GPT-Image-2 leads 9 of them, and GPT-Image-1.5 leads the other one. Arrows show how far a model moved from its overall rank among the models we track.
Following
7 models we track have a score here.
- 1 GPT-Image-1.5 ↑3 90.3%
- 2 GPT-Image-2 ↓1 90.1%
- 3
Gemini 3.1 Flash Image ↓1 88%
- 4
Seedream 4.0 ↓1 85.3%
- 5
Qwen-Image-Edit (2511) ↑1 47.7%
- 6 FLUX.2 [dev] ↓1 46.5%
- 7
BAGEL 37.3%
Consistency
7 models we track have a score here.
- 1 GPT-Image-2 87.4%
- 2
Gemini 3.1 Flash Image 82%
- 3
Seedream 4.0 73.4%
- 4 GPT-Image-1.5 72.1%
- 5 FLUX.2 [dev] 50%
- 6
Qwen-Image-Edit (2511) 46.4%
- 7
BAGEL 40.7%
Composition
7 models we track have a score here.
- 1 GPT-Image-2 90.8%
- 2 GPT-Image-1.5 ↑2 85.9%
- 3
Gemini 3.1 Flash Image ↓1 81.2%
- 4
Seedream 4.0 ↓1 80.7%
- 5
Qwen-Image-Edit (2511) ↑1 51.5%
- 6 FLUX.2 [dev] ↓1 51.1%
- 7
BAGEL 34.7%
Minimality
7 models we track have a score here.
- 1 GPT-Image-2 84.5%
- 2
Gemini 3.1 Flash Image 72.1%
- 3
Seedream 4.0 69.7%
- 4
BAGEL ↑3 58.1%
- 5 GPT-Image-1.5 ↓1 56.8%
- 6 FLUX.2 [dev] ↓1 51.5%
- 7
Qwen-Image-Edit (2511) ↓1 49.1%
Quality
7 models we track have a score here.
- 1 GPT-Image-2 93.2%
- 2 GPT-Image-1.5 ↑2 85.6%
- 3
Gemini 3.1 Flash Image ↓1 84.1%
- 4
Seedream 4.0 ↓1 82.8%
- 5 FLUX.2 [dev] 63.8%
- 6
Qwen-Image-Edit (2511) 61.3%
- 7
BAGEL 39.8%
Entity Fusion
7 models we track have a score here.
- 1 GPT-Image-2 91.5%
- 2
Gemini 3.1 Flash Image 82.5%
- 3 GPT-Image-1.5 ↑1 81.8%
- 4
Seedream 4.0 ↓1 77.9%
- 5 FLUX.2 [dev] 73.5%
- 6
Qwen-Image-Edit (2511) 66.9%
- 7
BAGEL 41.5%
Multi-Element Extraction
7 models we track have a score here.
- 1 GPT-Image-2 89.6%
- 2
Gemini 3.1 Flash Image 85.6%
- 3
Seedream 4.0 83.1%
- 4 GPT-Image-1.5 78.4%
- 5
Qwen-Image-Edit (2511) ↑1 55.3%
- 6 FLUX.2 [dev] ↓1 54.1%
- 7
BAGEL 39.9%
Visual Reasoning
7 models we track have a score here.
- 1 GPT-Image-2 84.8%
- 2
Seedream 4.0 ↑1 83.7%
- 3 GPT-Image-1.5 ↑1 76.5%
- 4
Gemini 3.1 Flash Image ↓2 75.9%
- 5
BAGEL ↑2 52.1%
- 6
Qwen-Image-Edit (2511) 44.3%
- 7 FLUX.2 [dev] ↓2 36.7%
Attribute Transfer
7 models we track have a score here.
- 1 GPT-Image-2 90.2%
- 2
Gemini 3.1 Flash Image 82.4%
- 3 GPT-Image-1.5 ↑1 82.3%
- 4
Seedream 4.0 ↓1 76.6%
- 5 FLUX.2 [dev] 47.9%
- 6
Qwen-Image-Edit (2511) 44.1%
- 7
BAGEL 35.6%
Infographic Manip.
7 models we track have a score here.
- 1 GPT-Image-2 87.6%
- 2
Gemini 3.1 Flash Image 79.8%
- 3 GPT-Image-1.5 ↑1 68.5%
- 4
Seedream 4.0 ↓1 67.9%
- 5 FLUX.2 [dev] 44%
- 6
BAGEL ↑1 40.5%
- 7
Qwen-Image-Edit (2511) ↓1 39.2%
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official MDMR-Bench 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
9 of 10 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 3 pass 1 note
- Pass Independent runs Scores from the official leaderboard (11 entries).
- Pass Models scored 7 models we track, plus 4 other entries on its leaderboard.
- 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.
- Pass Headroom The best is 81.2%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on github.com: 7 rows for models we track and 4 other entries 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.