Bio-VQA
Models identify organisms, describe visible traits, judge spatial relationships, and recognize individual animals across photographs. The benchmark has 24 tasks.
Passes test 2. The authors publish scores for 18 general vision-language models and release their individual answers. Each task has its own scoring rule.
- Made by
- Imageomics Institute and collaborators
- Released
- August 2026
- Kind
- General visual questions
- Level
- University
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Measures
- Capability
- Subject
- Biological sciences
- Language
- English
- Interaction
- Single turn
- Items from
- Another benchmark
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Vision, Biology and medicine
- Last model added
- June 2026
- Scores refreshed
- November 20th, 2026
InternVL3-38B leads with 0.91 among models released by June 2026 InternVL3-38B leads with 0.91 among models released by June 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 June 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
- 0.91
InternVL3-38B
- 0.89
Qwen3-VL-8B-Instruct
- 0.88
Gemma 4 12B
- 0.86 Qwen3-VL-30B-A3B
- 0.83
Qwen3-VL-32B-Instruct
- 0.83
InternVL3-8B
- 0.82
Molmo2-8B
- 0.82
InternVL3.5-8B
- 0.79
Gemma 3 12B
- 0.76
Qwen2.5-VL-7B
- 0.75 Phi-3.5-vision-instruct
- 0.75
InternVL3.5-38B
- 0.69
Pixtral 12B
- 0.66 llava-v1.6-vicuna-7b
- 0.55
Llama 3.2 11B
- 0.49
InternVL3.5-30B-A3B
- 0.42
Qwen2.5-VL-3B
- 0.36 llava-v1.6-mistral-7b
- 1
InternVL3-38B Zero-shot, greedy decoding 0.91
- 2
Qwen3-VL-8B-Instruct Zero-shot, greedy decoding 0.89
- 3
Gemma 4 12B Zero-shot, greedy decoding 0.88
- 4 Qwen3-VL-30B-A3B Zero-shot, greedy decoding 0.86
- 5
Qwen3-VL-32B-Instruct Zero-shot, greedy decoding 0.83
- 6
InternVL3-8B Zero-shot, greedy decoding 0.83
- 7
Molmo2-8B Zero-shot, greedy decoding 0.82
- 8
InternVL3.5-8B Zero-shot, greedy decoding 0.82
- 9
Gemma 3 12B Zero-shot, greedy decoding 0.79
- 10
Qwen2.5-VL-7B Zero-shot, greedy decoding 0.76
- 11 Phi-3.5-vision-instruct Zero-shot, greedy decoding 0.75
- 12
InternVL3.5-38B Zero-shot, greedy decoding 0.75
- 13
Pixtral 12B Zero-shot, greedy decoding 0.69
- 14 llava-v1.6-vicuna-7b Zero-shot, greedy decoding 0.66
- 15
Llama 3.2 11B Zero-shot, greedy decoding 0.55
- 16
InternVL3.5-30B-A3B Zero-shot, greedy decoding 0.49
- 17
Qwen2.5-VL-3B Zero-shot, greedy decoding 0.42
- 18 llava-v1.6-mistral-7b Zero-shot, greedy decoding 0.36
Source: Official Bio-VQA leaderboard
- 0.91
InternVL3-38B
- 0.89
Qwen3-VL-8B-Instruct
- 0.88
Gemma 4 12B
- 0.83
Qwen3-VL-32B-Instruct
- 0.83
InternVL3-8B
- 0.82
Molmo2-8B
- 0.82
InternVL3.5-8B
- 0.79
Gemma 3 12B
- 0.76
Qwen2.5-VL-7B
- 0.75 Phi-3.5-vision-instruct
- 0.75
InternVL3.5-38B
- 0.69
Pixtral 12B
- 0.66 llava-v1.6-vicuna-7b
- 0.55
Llama 3.2 11B
- 0.49
InternVL3.5-30B-A3B
- 0.42
Qwen2.5-VL-3B
- 0.36 llava-v1.6-mistral-7b
- 1
InternVL3-38B Zero-shot, greedy decoding 0.91
- 2
Qwen3-VL-8B-Instruct Zero-shot, greedy decoding 0.89
- 3
Gemma 4 12B Zero-shot, greedy decoding 0.88
- 4
Qwen3-VL-32B-Instruct Zero-shot, greedy decoding 0.83
- 5
InternVL3-8B Zero-shot, greedy decoding 0.83
- 6
Molmo2-8B Zero-shot, greedy decoding 0.82
- 7
InternVL3.5-8B Zero-shot, greedy decoding 0.82
- 8
Gemma 3 12B Zero-shot, greedy decoding 0.79
- 9
Qwen2.5-VL-7B Zero-shot, greedy decoding 0.76
- 10 Phi-3.5-vision-instruct Zero-shot, greedy decoding 0.75
- 11
InternVL3.5-38B Zero-shot, greedy decoding 0.75
- 12
Pixtral 12B Zero-shot, greedy decoding 0.69
- 13 llava-v1.6-vicuna-7b Zero-shot, greedy decoding 0.66
- 14
Llama 3.2 11B Zero-shot, greedy decoding 0.55
- 15
InternVL3.5-30B-A3B Zero-shot, greedy decoding 0.49
- 16
Qwen2.5-VL-3B Zero-shot, greedy decoding 0.42
- 17 llava-v1.6-mistral-7b Zero-shot, greedy decoding 0.36
Source: Official Bio-VQA leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 0.91 | Zero-shot, greedy decoding | Apr 2025 | Official leaderboard |
| 2 | | 0.89 | Zero-shot, greedy decoding | Oct 2025 | Official leaderboard |
| 3 | | 0.88 | Zero-shot, greedy decoding | Jun 2026 | Official leaderboard |
| 4 | Qwen3-VL-30B-A3B | 0.86 | Zero-shot, greedy decoding | Official leaderboard | |
| 5 | | 0.83 | Zero-shot, greedy decoding | Oct 2025 | Official leaderboard |
| 6 | | 0.83 | Zero-shot, greedy decoding | Apr 2025 | Official leaderboard |
| 7 | | 0.82 | Zero-shot, greedy decoding | Dec 2025 | Official leaderboard |
| 8 | | 0.82 | Zero-shot, greedy decoding | Aug 2025 | Official leaderboard |
| 9 | | 0.79 | Zero-shot, greedy decoding | Mar 2025 | Official leaderboard |
| 10 | | 0.76 | Zero-shot, greedy decoding | Jan 2025 | Official leaderboard |
| 11 | Phi-3.5-vision-instruct | 0.75 | Zero-shot, greedy decoding | Aug 2024 | Official leaderboard |
| 12 | | 0.75 | Zero-shot, greedy decoding | Aug 2025 | Official leaderboard |
| 13 | | 0.69 | Zero-shot, greedy decoding | Sep 2024 | Official leaderboard |
| 14 | llava-v1.6-vicuna-7b | 0.66 | Zero-shot, greedy decoding | Jan 2024 | Official leaderboard |
| 15 | | 0.55 | Zero-shot, greedy decoding | Sep 2024 | Official leaderboard |
| 16 | | 0.49 | Zero-shot, greedy decoding | Aug 2025 | Official leaderboard |
| 17 | | 0.42 | Zero-shot, greedy decoding | Jan 2025 | Official leaderboard |
| 18 | llava-v1.6-mistral-7b | 0.36 | Zero-shot, greedy decoding | Jan 2024 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 0.91 | Zero-shot, greedy decoding | Apr 2025 | Official leaderboard |
| 2 | | 0.89 | Zero-shot, greedy decoding | Oct 2025 | Official leaderboard |
| 3 | | 0.88 | Zero-shot, greedy decoding | Jun 2026 | Official leaderboard |
| 4 | | 0.83 | Zero-shot, greedy decoding | Oct 2025 | Official leaderboard |
| 5 | | 0.83 | Zero-shot, greedy decoding | Apr 2025 | Official leaderboard |
| 6 | | 0.82 | Zero-shot, greedy decoding | Dec 2025 | Official leaderboard |
| 7 | | 0.82 | Zero-shot, greedy decoding | Aug 2025 | Official leaderboard |
| 8 | | 0.79 | Zero-shot, greedy decoding | Mar 2025 | Official leaderboard |
| 9 | | 0.76 | Zero-shot, greedy decoding | Jan 2025 | Official leaderboard |
| 10 | Phi-3.5-vision-instruct | 0.75 | Zero-shot, greedy decoding | Aug 2024 | Official leaderboard |
| 11 | | 0.75 | Zero-shot, greedy decoding | Aug 2025 | Official leaderboard |
| 12 | | 0.69 | Zero-shot, greedy decoding | Sep 2024 | Official leaderboard |
| 13 | llava-v1.6-vicuna-7b | 0.66 | Zero-shot, greedy decoding | Jan 2024 | Official leaderboard |
| 14 | | 0.55 | Zero-shot, greedy decoding | Sep 2024 | Official leaderboard |
| 15 | | 0.49 | Zero-shot, greedy decoding | Aug 2025 | Official leaderboard |
| 16 | | 0.42 | Zero-shot, greedy decoding | Jan 2025 | Official leaderboard |
| 17 | llava-v1.6-mistral-7b | 0.36 | Zero-shot, greedy decoding | Jan 2024 | Official leaderboard |
Rankings by part
We read 22 parts from the benchmark’s own leaderboard in 4 groups: c1 taxonomy (exact match), c2 attributes (cags), c3 spatial and c4 re-id. Each ranks models on one part of the test. Qwen3-VL-32B-Instruct leads 12 of them, and 6 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
phylum
17 models we track have a score here.
- 1
InternVL3-38B 0.91
- 2
Qwen3-VL-8B-Instruct 0.89
- 3
Gemma 4 12B 0.88
- 4
Qwen3-VL-32B-Instruct 0.83
- 5
InternVL3-8B 0.83
- 6
Molmo2-8B 0.82
- 7
InternVL3.5-8B 0.82
- 8
Gemma 3 12B 0.79
- 9
Qwen2.5-VL-7B 0.76
- 10 Phi-3.5-vision-instruct 0.75
class
17 models we track have a score here.
- 1
Gemma 3 12B ↑7 0.69
- 2
Gemma 4 12B ↑1 0.69
- 3
InternVL3-38B ↓2 0.68
- 4
Qwen3-VL-8B-Instruct ↓2 0.68
- 5
InternVL3.5-38B ↑6 0.64
- 6
Qwen3-VL-32B-Instruct ↓2 0.61
- 7
Molmo2-8B ↓1 0.56
- 8
InternVL3.5-8B ↓1 0.53
- 9
Qwen2.5-VL-7B 0.50
- 10
Pixtral 12B ↑2 0.50
order
17 models we track have a score here.
- 1
Qwen3-VL-8B-Instruct ↑1 0.72
- 2
Qwen3-VL-32B-Instruct ↑2 0.71
- 3
Qwen2.5-VL-7B ↑6 0.67
- 4
InternVL3-38B ↓3 0.64
- 5
Gemma 3 12B ↑3 0.63
- 6
InternVL3.5-38B ↑5 0.62
- 7
Gemma 4 12B ↓4 0.58
- 8
InternVL3.5-30B-A3B ↑7 0.55
- 9
Pixtral 12B ↑3 0.53
- 10
Llama 3.2 11B ↑4 0.52
family
17 models we track have a score here.
- 1
Qwen3-VL-32B-Instruct ↑3 0.61
- 2
Qwen3-VL-8B-Instruct 0.59
- 3
Qwen2.5-VL-7B ↑6 0.54
- 4
InternVL3-38B ↓3 0.48
- 5
Qwen2.5-VL-3B ↑11 0.45
- 6
Llama 3.2 11B ↑8 0.43
- 7
Gemma 3 12B ↑1 0.43
- 8
InternVL3.5-38B ↑3 0.36
- 9
Gemma 4 12B ↓6 0.36
- 10
Pixtral 12B ↑2 0.35
genus
17 models we track have a score here.
- 1
Qwen3-VL-32B-Instruct ↑3 0.44
- 2
Qwen2.5-VL-7B ↑7 0.42
- 3
Qwen3-VL-8B-Instruct ↓1 0.41
- 4
Qwen2.5-VL-3B ↑12 0.36
- 5
InternVL3-38B ↓4 0.35
- 6
InternVL3.5-38B ↑5 0.27
- 7
InternVL3-8B ↓2 0.26
- 8
InternVL3.5-30B-A3B ↑7 0.24
- 9
Llama 3.2 11B ↑5 0.23
- 10
Gemma 3 12B ↓2 0.23
species
17 models we track have a score here.
- 1
Qwen3-VL-32B-Instruct ↑3 0.07
- 2
InternVL3-38B ↓1 0.05
- 3
Gemma 3 12B ↑5 0.05
- 4
Qwen2.5-VL-7B ↑5 0.05
- 5
Qwen3-VL-8B-Instruct ↓3 0.05
- 6
Qwen2.5-VL-3B ↑10 0.04
- 7
InternVL3.5-38B ↑4 0.04
- 8
InternVL3.5-30B-A3B ↑7 0.03
- 9
InternVL3-8B ↓4 0.03
- 10
Llama 3.2 11B ↑4 0.03
sci. name
17 models we track have a score here.
- 1
Qwen2.5-VL-7B ↑8 0.27
- 2
Qwen3-VL-32B-Instruct ↑2 0.24
- 3
Qwen2.5-VL-3B ↑13 0.21
- 4
Qwen3-VL-8B-Instruct ↓2 0.20
- 5
InternVL3-38B ↓4 0.19
- 6
InternVL3.5-38B ↑5 0.13
- 7
InternVL3-8B ↓2 0.12
- 8
InternVL3.5-30B-A3B ↑7 0.12
- 9
Llama 3.2 11B ↑5 0.10
- 10
Gemma 3 12B ↓2 0.09
common
17 models we track have a score here.
- 1
Qwen3-VL-32B-Instruct ↑3 0.08
- 2
Qwen3-VL-8B-Instruct 0.07
- 3
Qwen2.5-VL-3B ↑13 0.06
- 4
Qwen2.5-VL-7B ↑5 0.06
- 5
Llama 3.2 11B ↑9 0.05
- 6
Gemma 3 12B ↑2 0.05
- 7
InternVL3-38B ↓6 0.04
- 8
Gemma 4 12B ↓5 0.04
- 9
Pixtral 12B ↑3 0.03
- 10
InternVL3.5-38B ↑1 0.03
salience
17 models we track have a score here.
- 1
Gemma 4 12B ↑2 0.59
- 2
Qwen3-VL-32B-Instruct ↑2 0.58
- 3
InternVL3-38B ↓2 0.56
- 4
InternVL3.5-38B ↑7 0.56
- 5
Qwen2.5-VL-7B ↑4 0.54
- 6
InternVL3.5-30B-A3B ↑9 0.54
- 7
Qwen3-VL-8B-Instruct ↓5 0.53
- 8
InternVL3-8B ↓3 0.51
- 9
Molmo2-8B ↓3 0.50
- 10
InternVL3.5-8B ↓3 0.49
morphology
17 models we track have a score here.
- 1
InternVL3-38B 0.65
- 2
Qwen3-VL-32B-Instruct ↑2 0.64
- 3
Molmo2-8B ↑3 0.64
- 4
InternVL3.5-38B ↑7 0.64
- 5
Qwen3-VL-8B-Instruct ↓3 0.62
- 6
Gemma 4 12B ↓3 0.62
- 7
InternVL3.5-30B-A3B ↑8 0.61
- 8
InternVL3.5-8B ↓1 0.60
- 9
InternVL3-8B ↓4 0.58
- 10
Qwen2.5-VL-7B ↓1 0.58
localization
17 models we track have a score here.
- 1
Qwen3-VL-32B-Instruct ↑3 0.47
- 2
Molmo2-8B ↑4 0.47
- 3
Qwen3-VL-8B-Instruct ↓1 0.44
- 4
Gemma 4 12B ↓1 0.43
- 5
InternVL3-38B ↓4 0.41
- 6
Gemma 3 12B ↑2 0.41
- 7
Qwen2.5-VL-3B ↑9 0.39
- 8
InternVL3.5-38B ↑3 0.38
- 9
InternVL3.5-8B ↓2 0.37
- 10 llava-v1.6-vicuna-7b ↑3 0.36
naming
17 models we track have a score here.
- 1
Qwen3-VL-32B-Instruct ↑3 0.73
- 2
Qwen2.5-VL-7B ↑7 0.71
- 3
InternVL3-38B ↓2 0.67
- 4
Qwen2.5-VL-3B ↑12 0.66
- 5
Qwen3-VL-8B-Instruct ↓3 0.64
- 6
InternVL3-8B ↓1 0.58
- 7
InternVL3.5-30B-A3B ↑8 0.54
- 8
InternVL3.5-38B ↑3 0.54
- 9
Llama 3.2 11B ↑5 0.52
- 10
Gemma 3 12B ↓2 0.51
count
17 models we track have a score here.
- 1
InternVL3-38B 0.80
- 2
Qwen3-VL-32B-Instruct ↑2 0.79
- 3
InternVL3.5-38B ↑8 0.77
- 4
Gemma 4 12B ↓1 0.75
- 5
InternVL3.5-8B ↑2 0.75
- 6
Qwen3-VL-8B-Instruct ↓4 0.74
- 7
InternVL3-8B ↓2 0.74
- 8
InternVL3.5-30B-A3B ↑7 0.73
- 9
Llama 3.2 11B ↑5 0.72
- 10
Qwen2.5-VL-7B ↓1 0.72
rel. count
17 models we track have a score here.
- 1
InternVL3-38B 0.57
- 2
Gemma 4 12B ↑1 0.56
- 3
Qwen3-VL-32B-Instruct ↑1 0.56
- 4
InternVL3.5-38B ↑7 0.50
- 5
InternVL3.5-30B-A3B ↑10 0.49
- 6
Molmo2-8B 0.42
- 7
Gemma 3 12B ↑1 0.41
- 8
Qwen3-VL-8B-Instruct ↓6 0.41
- 9
Qwen2.5-VL-7B 0.41
- 10
InternVL3.5-8B ↓3 0.41
abs. dist.
17 models we track have a score here.
- 1
Qwen3-VL-32B-Instruct ↑3 0.62
- 2
InternVL3.5-38B ↑9 0.50
- 3
Gemma 4 12B 0.39
- 4
InternVL3.5-30B-A3B ↑11 0.38
- 5
InternVL3-38B ↓4 0.37
- 6
Gemma 3 12B ↑2 0.35
- 7
Qwen2.5-VL-7B ↑2 0.33
- 8 llava-v1.6-mistral-7b ↑9 0.33
- 9
Pixtral 12B ↑3 0.31
- 10
Qwen3-VL-8B-Instruct ↓8 0.30
rel. dist.
17 models we track have a score here.
- 1
Qwen3-VL-32B-Instruct ↑3 0.44
- 2
Gemma 4 12B ↑1 0.44
- 3
InternVL3-38B ↓2 0.36
- 4
InternVL3.5-38B ↑7 0.36
- 5
Molmo2-8B ↑1 0.36
- 6
Qwen3-VL-8B-Instruct ↓4 0.34
- 7
Qwen2.5-VL-7B ↑2 0.34
- 8
InternVL3.5-30B-A3B ↑7 0.33
- 9
Llama 3.2 11B ↑5 0.32
- 10
Pixtral 12B ↑2 0.30
localize
17 models we track have a score here.
- 1
Qwen3-VL-32B-Instruct ↑3 0.87
- 2
Qwen3-VL-8B-Instruct 0.86
- 3
InternVL3-38B ↓2 0.74
- 4 llava-v1.6-vicuna-7b ↑9 0.70
- 5
Gemma 4 12B ↓2 0.59
- 6
InternVL3.5-38B ↑5 0.41
- 7
Gemma 3 12B ↑1 0.33
- 8
Qwen2.5-VL-7B ↑1 0.31
- 9
Qwen2.5-VL-3B ↑7 0.31
- 10
InternVL3.5-30B-A3B ↑5 0.29
rel. pos.
17 models we track have a score here.
- 1
Qwen3-VL-32B-Instruct ↑3 0.55
- 2
Gemma 4 12B ↑1 0.49
- 3
Molmo2-8B ↑3 0.46
- 4
Qwen3-VL-8B-Instruct ↓2 0.42
- 5
Pixtral 12B ↑7 0.27
- 6
InternVL3.5-30B-A3B ↑9 0.25
- 7
InternVL3.5-8B 0.22
- 8
InternVL3.5-38B ↑3 0.22
- 9
Qwen2.5-VL-7B 0.21
- 10
InternVL3-38B ↓9 0.20
viewpoint
17 models we track have a score here.
- 1
Molmo2-8B ↑5 0.48
- 2
Qwen3-VL-32B-Instruct ↑2 0.43
- 3
Qwen2.5-VL-7B ↑6 0.40
- 4
InternVL3-38B ↓3 0.37
- 5
Qwen3-VL-8B-Instruct ↓3 0.37
- 6
InternVL3.5-38B ↑5 0.34
- 7
InternVL3-8B ↓2 0.34
- 8
Gemma 4 12B ↓5 0.33
- 9
InternVL3.5-30B-A3B ↑6 0.31
- 10
InternVL3.5-8B ↓3 0.26
OCR
17 models we track have a score here.
- 1
Qwen2.5-VL-7B ↑8 0.92
- 2
Qwen2.5-VL-3B ↑14 0.90
- 3
Qwen3-VL-32B-Instruct ↑1 0.90
- 4
Llama 3.2 11B ↑10 0.72
- 5
Qwen3-VL-8B-Instruct ↓3 0.71
- 6
Molmo2-8B 0.69
- 7
InternVL3.5-8B 0.46
- 8
Gemma 3 12B 0.35
- 9
Gemma 4 12B ↓6 0.34
- 10
InternVL3-8B ↓5 0.31
d′
17 models we track have a score here.
- 1
Qwen3-VL-32B-Instruct ↑3 0.70
- 2
InternVL3-8B ↑3 0.68
- 3
Qwen2.5-VL-3B ↑13 0.66
- 4
InternVL3.5-30B-A3B ↑11 0.57
- 5 Phi-3.5-vision-instruct ↑5 0.56
- 6
Qwen3-VL-8B-Instruct ↓4 0.49
- 7
Molmo2-8B ↓1 0.49
- 8 llava-v1.6-mistral-7b ↑9 0.48
- 9
InternVL3-38B ↓8 0.46
- 10
Qwen2.5-VL-7B ↓1 0.44
acc
17 models we track have a score here.
- 1
Qwen3-VL-32B-Instruct ↑3 0.64
- 2
InternVL3.5-30B-A3B ↑13 0.61
- 3
InternVL3-8B ↑2 0.60
- 4
InternVL3-38B ↓3 0.58
- 5
Molmo2-8B ↑1 0.58
- 6 Phi-3.5-vision-instruct ↑4 0.58
- 7
Qwen3-VL-8B-Instruct ↓5 0.57
- 8
Gemma 4 12B ↓5 0.57
- 9
Qwen2.5-VL-3B ↑7 0.56
- 10
InternVL3.5-8B ↓3 0.54
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official Bio-VQA leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Bio-VQA 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
All 9 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (18 entries).
- Pass Models scored 17 models we track, plus 1 other entry on its leaderboard.
- Pass Newest score The newest is from June 2026, for Gemma 4 12B.
- 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 smrayeed.github.io: 17 rows for models we track and 1 other entry 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.