InfiHall-Bench
Models generate images from 411 prompts that combine visual requirements. The evaluator checks details such as object parts, actions, and spatial relations separately.
- Made by
- Harbin Institute of Technology, Shenzhen and collaborators
- Released
- November 2026
- Kind
- Image generation and editing
- Level
- Everyday
- Scored
- Graded by a model
- Answers
- Open-ended
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- A host
- In views
- Image, video and 3D generation
- Last model added
- January 2025
- Scores refreshed
- November 20th, 2026
Infinity-8B tops the leaderboard with 67.5 FLUX.1 [dev] leads with 67.1 among models released by January 2025
Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. 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 FLUX.1 [dev], at 67.1. This view shows only models we track, each at its best.
The newest model with a score here came out in January 2025, 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 Infinity-8B CMHM 67.5
- 2 FLUX.1 [dev] 67.1
- 3 Infinity-8B-512 CMHM 64.8
- 4
Stable Diffusion 3.5 Large 64.0
- 5 SD3.5-Medium 61.7
- 6 Infinity-2B CMHM 58.1
- 7 PixArt-Sigma-XL-2-1024-MS 51.8
- 8 SDXL 47.5
- 9
Janus-Pro-7B 46.7
- 10 LlamaGen 34.3
Source: The InfiHall-Bench paper
- 1 FLUX.1 [dev] 67.1
- 2
Stable Diffusion 3.5 Large 64.0
- 3
Janus-Pro-7B 46.7
Source: The InfiHall-Bench paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Infinity-8B | 67.5 | CMHM | Official leaderboard | |
| 2 | FLUX.1 [dev] | 67.1 | Official leaderboard | ||
| 3 | Infinity-8B-512 | 64.8 | CMHM | Official leaderboard | |
| 4 | | 64.0 | Official leaderboard | ||
| 5 | SD3.5-Medium | 61.7 | Official leaderboard | ||
| 6 | Infinity-2B | 58.1 | CMHM | Official leaderboard | |
| 7 | PixArt-Sigma-XL-2-1024-MS | 51.8 | Official leaderboard | ||
| 8 | SDXL | 47.5 | Official leaderboard | ||
| 9 | | 46.7 | Jan 2025 | Official leaderboard | |
| 10 | LlamaGen | 34.3 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | FLUX.1 [dev] | 67.1 | Official leaderboard | ||
| 2 | | 64.0 | Official leaderboard | ||
| 3 | | 46.7 | Jan 2025 | Official leaderboard |
Rankings by part
We read 5 parts from the benchmark’s paper, each ranking models on one part of the test, such as Action Rel., Attribute and Part. FLUX.1 [dev] leads 4 of them, and Stable Diffusion 3.5 Large leads the other one. Arrows show how far a model moved from its overall rank among the models we track.
Action Rel.
3 models we track have a score here.
- 1 FLUX.1 [dev] 66.8
- 2
Stable Diffusion 3.5 Large 61.9
- 3
Janus-Pro-7B 37.3
Attribute
3 models we track have a score here.
- 1 FLUX.1 [dev] 90.3
- 2
Stable Diffusion 3.5 Large 86.6
- 3
Janus-Pro-7B 66.4
Part
3 models we track have a score here.
- 1 FLUX.1 [dev] 75.7
- 2
Stable Diffusion 3.5 Large 75.6
- 3
Janus-Pro-7B 55.1
Scene
3 models we track have a score here.
- 1
Stable Diffusion 3.5 Large ↑1 93.8
- 2 FLUX.1 [dev] ↓1 92.9
- 3
Janus-Pro-7B 78.3
Spatial Rel.
3 models we track have a score here.
- 1 FLUX.1 [dev] 54.1
- 2
Stable Diffusion 3.5 Large 53.3
- 3
Janus-Pro-7B 43.3
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official InfiHall-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
7 of 9 checks pass, 2 warnings. 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 1 warning 2 pass 2 notes
- Pass Independent runs Scores from the benchmark’s paper (10 entries).
- Pass Models scored 3 models we track, plus 7 other entries on its leaderboard.
- Warning Newest score The newest came 22 months ago, in January 2025, for Janus-Pro-7B.
- 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 arxiv.org: 3 rows for models we track and 7 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.