MATH-Vision (MathVision)
The official leaderboard hasn’t scored a model released after Seed2.0 Lite (0428) (April 2026). Labs report their own numbers for 14 newer models, such as Qwen 3.8 Max at 97.7%.
3,040 competition math problems with visual context. Answer accuracy.
- Made by
- CUHK MMLab
- Released
- February 2024
- Kind
- Visual math and diagrams
- Level
- Competition
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Measures
- Capability
- Subject
- Mathematics
- Language
- English
- Interaction
- Single turn
- Items from
- Exams and contests
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Vision
- Strongest model tested
- Kimi K3 #25 on the capability index
- Last in a launch post
- November 2026
- Last model added
- November 2026
- Scores refreshed
- November 20th, 2026
In the authors’ words
Recent advancements in Large Multimodal Models (LMMs) have shown promising results in mathematical reasoning within visual contexts, with models approaching human-level performance on existing benchmarks such as MathVista. However, we observe significant limitations in the diversity of questions and breadth of subjects covered by these benchmarks. To address this issue, we present the MATH-Vision (MATH-V) dataset, a meticulously curated collection of 3,040 high-quality mathematical problems with visual contexts sourced from real math competitions. Spanning 16 distinct mathematical disciplines and graded across 5 levels of difficulty, our dataset provides a comprehensive and diverse set of challenges for evaluating the mathematical reasoning abilities of LMMs. Through extensive experimentation, we unveil a notable performance gap between current LMMs and human performance on MATH-V, underscoring the imperative for further advancements in LMMs. Moreover, our detailed categorization allows for a thorough error analysis of LMMs, offering valuable insights to guide future research and development. The project is available at https://mathvision-cuhk.github.io
Qwen 3.8 Max leads with 97.7%
Best score per model as labs report them in their own cards, drawn striped. Labs pick their own settings, so these compare less cleanly than independent runs.
GPT-5.4 leads with 96.1% across all independent runs
Best score per model across the official leaderboard, Epoch AI’s own runs and any board Epoch AI copied. Dotted bars are reruns by Epoch AI or another board, and each row names its source.
Qwen 3.8 Max leads with 97.7%, claims included
Best score per model from every source, including labs’ own numbers and claims Epoch AI collected from reports and papers. Labs pick their own settings, so a claim can beat an independent run of the same model. Striped bars are claims, and dotted bars are reruns by Epoch AI or another board.
- 97.7
Qwen 3.8 Max
- 96.5
LongCat-2.5
- 96.2
Qwen3.8-Omni-Flash
- 95.7
Qwen3.8 Flash
- 94.6
Qwen 3.8 27B
- 94.5
Seed2.1 Pro
- 94.3 Kimi K3
- 92.8
Inkling 2 Small
- 90.3
Qwen3.7-Plus
- 89.8
Seed2.0 Lite (0428)
- 89.0
Inkling 2 Mini
- 88.8
Seed2.0 Pro
- 88.6
Qwen3.5 397B-A17B
- 88.0
Qwen 3.6 Plus
- 87.4 Kimi K2.6
- 86.2
Qwen3.5-122B-A10B
- 86.0
Qwen3.5-27B
- 85.6
Gemma 4 31B IT
- 84.2 Kimi K2.5
- 83.9
Qwen3.5-35B-A3B
- 82.4
Gemma 4 26B A4B
- 81.3
Seed1.8
- 79.7
Gemma 4 12B
- 78.9
Qwen3.5-9B
- 78.2
Nemotron 3.5 Super VL
- 75.2
EXAONE 4.5 33B
- 74.6
Qwen3-VL-235B-A22B
- 74.6
Qwen3.5-4B
- 74.3
ERNIE 5.0
- 74.2
MiniCPM-V-4.7-35B-A3B
- 73.0
Qwen3.5-Omni-Plus
- 70.8
STEP3-VL-10B
- 70.5
DiffusionGemma 26B A4B
- 70.3
Seed1.6-Thinking
- 68.7
Seed1.5-VL
- 65.6 GLM-4.5V
- 64.8
Step 3
- 63.9
InternVL3.5-241B-A28B
- 63.5 GLM-4.6V
- 62.5
Intern-S1
- 61.9
ERNIE-4.5-VL-28B-A3B-Thinking
- 59.5
Gemma 4 E4B
- 56.9 Kimi-VL-A3B-Thinking-2506
- 56.3
Qwen3-Omni-30B-A3B
- 54.4 GLM-4.1V-9B-Thinking
- 53.6
Nemotron Nano 12B v2 VL
- 52.4
Gemma 4 E2B
- 46.3
Nemotron 3.5 Bolt E4B
- 38.6 Kimi k1.5
- 38.1
Qwen2.5-VL-72B
- 36.8 Kimi-VL-A3B-Thinking
- 36.2
Phi-4-Reasoning-Vision-15B
- 25.0
Qwen2.5-Omni-7B
- 1
Qwen 3.8 Max 97.7%
- 2
LongCat-2.5 96.5%
- 3
Qwen3.8-Omni-Flash 96.2%
- 4
Qwen3.8 Flash 95.7%
- 5
Qwen 3.8 27B 94.6%
- 6
Seed2.1 Pro 94.5%
- 7 Kimi K3 94.3%
- 8
Inkling 2 Small 92.8%
- 9
Qwen3.7-Plus 90.3%
- 10
Seed2.0 Lite (0428) 89.8%
- 11
Inkling 2 Mini 89%
- 12
Seed2.0 Pro 88.8%
- 13
Qwen3.5 397B-A17B 88.6%
- 14
Qwen 3.6 Plus 88%
- 15 Kimi K2.6 87.4%
- 16
Qwen3.5-122B-A10B 86.2%
- 17
Qwen3.5-27B 86%
- 18
Gemma 4 31B IT 85.6%
- 19 Kimi K2.5 84.2%
- 20
Qwen3.5-35B-A3B 83.9%
- 21
Gemma 4 26B A4B 82.4%
- 22
Seed1.8 81.3%
- 23
Gemma 4 12B 79.7%
- 24
Qwen3.5-9B 78.9%
- 25
Nemotron 3.5 Super VL 78.2%
- 26
EXAONE 4.5 33B 75.2%
- 27
Qwen3-VL-235B-A22B 74.6%
- 28
Qwen3.5-4B 74.6%
- 29
ERNIE 5.0 74.3%
- 30
MiniCPM-V-4.7-35B-A3B 74.2%
- 31
Qwen3.5-Omni-Plus 73%
- 32
STEP3-VL-10B 70.8%
- 33
DiffusionGemma 26B A4B 70.5%
- 34
Seed1.6-Thinking 70.3%
- 35
Seed1.5-VL 68.7%
- 36 GLM-4.5V 65.6%
- 37
Step 3 64.8%
- 38
InternVL3.5-241B-A28B 63.9%
- 39 GLM-4.6V 63.5%
- 40
Intern-S1 62.5%
- 41
ERNIE-4.5-VL-28B-A3B-Thinking 61.9%
- 42
Gemma 4 E4B 59.5%
- 43 Kimi-VL-A3B-Thinking-2506 56.9%
- 44
Qwen3-Omni-30B-A3B 56.3%
- 45 GLM-4.1V-9B-Thinking 54.4%
- 46
Nemotron Nano 12B v2 VL 53.6%
- 47
Gemma 4 E2B 52.4%
- 48
Nemotron 3.5 Bolt E4B 46.3%
- 49 Kimi k1.5 38.6%
- 50
Qwen2.5-VL-72B 38.1%
- 51 Kimi-VL-A3B-Thinking 36.8%
- 52
Phi-4-Reasoning-Vision-15B 36.2%
- 53
Qwen2.5-Omni-7B 25%
Source: Labs’ model cards
- 96.1 GPT-5.4
- 95.7
Gemini 3.1 Pro
- 93.2 Kimi K2.6
- 90.6
Gemini 3 Flash
- 89.8
Seed2.0 Lite (0428)
- 88.8
Seed2.0 Pro
- 88.6
Qwen3.5 397B-A17B
- 86.6
Gemini 3 Pro
- 86.4
Seed2.0 Lite (0215)
- 86.2
Qwen3.5-122B-A10B
- 86.0
Qwen3.5-27B
- 85.6
Gemma 4 31B IT
- 85.0 Kimi K2.5
- 84.6 Claude Opus 4.6
- 83.9
Qwen3.5-35B-A3B
- 83.0 GPT-5.2
- 82.4
Gemma 4 26B A4B
- 81.3
Seed1.8
- 77.2
Seed1.6-Thinking
- 77.1 Claude Opus 4.5
- 74.6
Qwen3-VL-235B-A22B
- 73.3
Gemini 2.5 Pro
- 72.0 GPT-5
- 71.9 GPT-5 mini
- 71.1 Claude Sonnet 4.5
- 69.6 dots.vlm1
- 68.7
Seed1.5-VL
- 66.0 Claude Opus 4.1
- 66.0
Qwen3-VL-235B-A22B-Instruct
- 65.6 GLM-4.5V
- 64.8
Step 3
- 63.9
InternVL3.5-241B-A28B
- 63.5 GLM-4.6V
- 60.4 MiMo-VL-7B-RL
- 60.3 o1
- 59.6 MiMo-VL-7B-RL-2508
- 59.6
Qwen3-VL-8B-Thinking
- 59.5
Gemma 4 E4B
- 58.6 Claude 3.7 Sonnet
- 58.0 o4-mini
- 57.9 MiMo-VL-7B-SFT
- 56.9 Kimi-VL-A3B-Thinking-2506
- 55.7
InternVL3.5-30B-A3B
- 54.4 GLM-4.1V-9B-Thinking
- 54.0 GLM-4.6V-Flash
- 54.0
InternVL3.5-38B
- 53.9
Ovis2.5-9B
- 52.6
Skywork-R1V3-38B
- 52.4
Gemma 4 E2B
- 52.0
InternVL3.5-8B
- 50.3
Qianfan-VL-70B
- 49.7
Skywork-R1V2-38B
- 48.6
Doubao-1.5-pro
- 48.1
Gemini 2.0 Pro
- 47.3 GPT-4.5
- 46.0
Gemma 3 27B
- 46.0
Keye-VL-8B-Preview
- 43.1
InternVL3-78B
- 41.3
Gemini 2.0 Flash
- 40.0
InternVL3.5-4B
- 38.6 Kimi k1.5
- 38.1
Qwen2.5-VL-72B
- 38.0 Claude 3.5 Sonnet
- 37.4
Ovis2.5-2B
- 37.0
InternVL3-14B
- 36.8 Kimi-VL-A3B-Thinking
- 35.9
QVQ-72B-Preview
- 34.5
InternVL3-38B
- 32.8
Qianfan-VL-8B
- 32.2
InternVL2_5-78B
- 31.9
Ovis2-34B
- 31.8
InternVL2.5-38B
- 30.4 GPT-4o
- 30.3 GPT-4 Turbo
- 30.1
Ovis2-16B
- 29.0
InternVL3-8B
- 27.6
InternVL3-9B
- 27.1 Claude 3 Opus
- 25.9 Qwen2-VL-72B-Instruct
- 25.3 LLaVA-OneVision 72B
- 25.1
Qwen2.5-VL-7B
- 23.4 MiniCPM-V-2_6
- 23.1
InternVL2.5-26B
- 22.8 GPT-4V
- 21.7 MiniCPM-o-2_6
- 21.7
InternVL3-2B
- 21.2
Qwen2.5-VL-3B
- 19.7
InternVL2.5-8B
- 19.2
Gemini 1.5 Pro
- 18.8
InternVL3-1B
- 18.8
Ovis1.6-Gemma2-9B
- 17.9 Aquila-VL-2B
- 17.7
Gemini 1.0 Pro
- 17.0
InternVL-Chat-V1-2-Plus
- 16.3 Qwen2-VL-7B-Instruct
- 15.6 Qwen-VL-Max
- 14.7
InternVL2.5-2B
- 14.5
InternLM-XComposer2-VL-7B
- 14.4
InternVL2.5-1B
- 12.4
Qwen2-VL-2B-Instruct
- 11.9
ShareGPT4V-13B
- 11.1
LLaVA-v1.5-13B
- 10.7
Qwen-VL-Plus
- 10.5
ShareGPT4V-7B
- 8.52 llava-v1.5-7b
- 1 GPT-5.4 xhigh, with python 96.1%
- 2
Gemini 3.1 Pro thinking high, with python 95.7%
- 3 Kimi K2.6 with python 93.2%
- 4
Gemini 3 Flash 90.6%
- 5
Seed2.0 Lite (0428) 89.8%
- 6
Seed2.0 Pro 88.8%
- 7
Qwen3.5 397B-A17B 88.6%
- 8
Gemini 3 Pro 86.6%
- 9
Seed2.0 Lite (0215) 86.4%
- 10
Qwen3.5-122B-A10B 86.2%
- 11
Qwen3.5-27B 86%
- 12
Gemma 4 31B IT 85.6%
- 13 Kimi K2.5 thinking, with python 85%
- 14 Claude Opus 4.6 max, with python 84.6%
- 15
Qwen3.5-35B-A3B 83.9%
- 16 GPT-5.2 83%
- 17
Gemma 4 26B A4B 82.4%
- 18
Seed1.8 81.3%
- 19
Seed1.6-Thinking 77.2%
- 20 Claude Opus 4.5 77.1%
- 21
Qwen3-VL-235B-A22B thinking 74.6%
- 22
Gemini 2.5 Pro Mar 2025 73.3%
- 23 GPT-5 72%
- 24 GPT-5 mini 71.9%
- 25 Claude Sonnet 4.5 71.1%
- 26 dots.vlm1 69.6%
- 27
Seed1.5-VL 68.7%
- 28 Claude Opus 4.1 thinking 66%
- 29
Qwen3-VL-235B-A22B-Instruct 66%
- 30 GLM-4.5V 65.6%
- 31
Step 3 64.8%
- 32
InternVL3.5-241B-A28B 63.9%
- 33 GLM-4.6V 63.5%
- 34 MiMo-VL-7B-RL 60.4%
- 35 o1 60.3%
- 36 MiMo-VL-7B-RL-2508 59.6%
- 37
Qwen3-VL-8B-Thinking 59.6%
- 38
Gemma 4 E4B 59.5%
- 39 Claude 3.7 Sonnet 58.6%
- 40 o4-mini 58%
- 41 MiMo-VL-7B-SFT 57.9%
- 42 Kimi-VL-A3B-Thinking-2506 56.9%
- 43
InternVL3.5-30B-A3B 55.7%
- 44 GLM-4.1V-9B-Thinking 54.4%
- 45 GLM-4.6V-Flash 54%
- 46
InternVL3.5-38B 54%
- 47
Ovis2.5-9B 53.9%
- 48
Skywork-R1V3-38B 52.6%
- 49
Gemma 4 E2B 52.4%
- 50
InternVL3.5-8B 52%
- 51
Qianfan-VL-70B 50.3%
- 52
Skywork-R1V2-38B 49.7%
- 53
Doubao-1.5-pro 48.6%
- 54
Gemini 2.0 Pro 48.1%
- 55 GPT-4.5 47.3%
- 56
Gemma 3 27B non-thinking 46%
- 57
Keye-VL-8B-Preview 46%
- 58
InternVL3-78B 43.1%
- 59
Gemini 2.0 Flash Feb 2025 41.3%
- 60
InternVL3.5-4B 40%
- 61 Kimi k1.5 38.6%
- 62
Qwen2.5-VL-72B 38.1%
- 63 Claude 3.5 Sonnet 38%
- 64
Ovis2.5-2B 37.4%
- 65
InternVL3-14B 37%
- 66 Kimi-VL-A3B-Thinking 36.8%
- 67
QVQ-72B-Preview 35.9%
- 68
InternVL3-38B 34.5%
- 69
Qianfan-VL-8B 32.8%
- 70
InternVL2_5-78B 32.2%
- 71
Ovis2-34B 31.9%
- 72
InternVL2.5-38B 31.8%
- 73 GPT-4o May 2024 30.4%
- 74 GPT-4 Turbo Apr 2024 30.3%
- 75
Ovis2-16B 30.1%
- 76
InternVL3-8B 29%
- 77
InternVL3-9B 27.6%
- 78 Claude 3 Opus 27.1%
- 79 Qwen2-VL-72B-Instruct 25.9%
- 80 LLaVA-OneVision 72B 25.3%
- 81
Qwen2.5-VL-7B 25.1%
- 82 MiniCPM-V-2_6 23.4%
- 83
InternVL2.5-26B 23.1%
- 84 GPT-4V 22.8%
- 85 MiniCPM-o-2_6 21.7%
- 86
InternVL3-2B 21.7%
- 87
Qwen2.5-VL-3B 21.2%
- 88
InternVL2.5-8B 19.7%
- 89
Gemini 1.5 Pro May 2024 19.2%
- 90
InternVL3-1B 18.8%
- 91
Ovis1.6-Gemma2-9B 18.8%
- 92 Aquila-VL-2B 17.9%
- 93
Gemini 1.0 Pro 17.7%
- 94
InternVL-Chat-V1-2-Plus 17%
- 95 Qwen2-VL-7B-Instruct 16.3%
- 96 Qwen-VL-Max 15.6%
- 97
InternVL2.5-2B 14.7%
- 98
InternLM-XComposer2-VL-7B 14.5%
- 99
InternVL2.5-1B 14.4%
- 100
Qwen2-VL-2B-Instruct 12.4%
- 101
ShareGPT4V-13B 11.9%
- 102
LLaVA-v1.5-13B 11.1%
- 103
Qwen-VL-Plus 10.7%
- 104
ShareGPT4V-7B 10.5%
- 105 llava-v1.5-7b 8.52%
Source: Official MATH-Vision (MathVision) leaderboard
- 97.7
Qwen 3.8 Max
- 96.5
LongCat-2.5
- 96.2
Qwen3.8-Omni-Flash
- 96.1 GPT-5.4
- 95.7
Gemini 3.1 Pro
- 95.7
Qwen3.8 Flash
- 94.6
Qwen 3.8 27B
- 94.5
Seed2.1 Pro
- 94.3 Kimi K3
- 93.2 Kimi K2.6
- 92.8
Inkling 2 Small
- 90.6
Gemini 3 Flash
- 90.3
Qwen3.7-Plus
- 89.8
Seed2.0 Lite (0428)
- 89.0
Inkling 2 Mini
- 88.8
Seed2.0 Pro
- 88.6
Qwen3.5 397B-A17B
- 88.0
Qwen 3.6 Plus
- 86.6
Gemini 3 Pro
- 86.4
Seed2.0 Lite (0215)
- 86.2
Qwen3.5-122B-A10B
- 86.0
Qwen3.5-27B
- 85.6
Gemma 4 31B IT
- 85.0 Kimi K2.5
- 84.6 Claude Opus 4.6
- 83.9
Qwen3.5-35B-A3B
- 83.0 GPT-5.2
- 82.4
Gemma 4 26B A4B
- 81.3
Seed1.8
- 79.7
Gemma 4 12B
- 78.9
Qwen3.5-9B
- 78.2
Nemotron 3.5 Super VL
- 77.2
Seed1.6-Thinking
- 77.1 Claude Opus 4.5
- 75.2
EXAONE 4.5 33B
- 74.6
Qwen3-VL-235B-A22B
- 74.6
Qwen3.5-4B
- 74.3
ERNIE 5.0
- 74.2
MiniCPM-V-4.7-35B-A3B
- 73.3
Gemini 2.5 Pro
- 73.0
Qwen3.5-Omni-Plus
- 72.0 GPT-5
- 71.9 GPT-5 mini
- 71.1 Claude Sonnet 4.5
- 70.8
STEP3-VL-10B
- 70.5
DiffusionGemma 26B A4B
- 69.6 dots.vlm1
- 68.7
Seed1.5-VL
- 66.0 Claude Opus 4.1
- 66.0
Qwen3-VL-235B-A22B-Instruct
- 65.6 GLM-4.5V
- 64.8
Step 3
- 63.9
InternVL3.5-241B-A28B
- 63.5 GLM-4.6V
- 62.5
Intern-S1
- 61.9
ERNIE-4.5-VL-28B-A3B-Thinking
- 60.4 MiMo-VL-7B-RL
- 60.3 o1
- 59.6 MiMo-VL-7B-RL-2508
- 59.6
Qwen3-VL-8B-Thinking
- 59.5
Gemma 4 E4B
- 58.6 Claude 3.7 Sonnet
- 58.0 o4-mini
- 57.9 MiMo-VL-7B-SFT
- 56.9 Kimi-VL-A3B-Thinking-2506
- 56.3
Qwen3-Omni-30B-A3B
- 55.7
InternVL3.5-30B-A3B
- 54.4 GLM-4.1V-9B-Thinking
- 54.0 GLM-4.6V-Flash
- 54.0
InternVL3.5-38B
- 53.9
Ovis2.5-9B
- 53.6
Nemotron Nano 12B v2 VL
- 52.6
Skywork-R1V3-38B
- 52.4
Gemma 4 E2B
- 52.0
InternVL3.5-8B
- 50.3
Qianfan-VL-70B
- 49.7
Skywork-R1V2-38B
- 48.6
Doubao-1.5-pro
- 48.1
Gemini 2.0 Pro
- 47.3 GPT-4.5
- 46.3
Nemotron 3.5 Bolt E4B
- 46.0
Gemma 3 27B
- 46.0
Keye-VL-8B-Preview
- 43.1
InternVL3-78B
- 41.3
Gemini 2.0 Flash
- 40.0
InternVL3.5-4B
- 38.6 Kimi k1.5
- 38.1
Qwen2.5-VL-72B
- 38.0 Claude 3.5 Sonnet
- 37.4
Ovis2.5-2B
- 37.0
InternVL3-14B
- 36.8 Kimi-VL-A3B-Thinking
- 36.2
Phi-4-Reasoning-Vision-15B
- 35.9
QVQ-72B-Preview
- 34.5
InternVL3-38B
- 32.8
Qianfan-VL-8B
- 32.2
InternVL2_5-78B
- 31.9
Ovis2-34B
- 31.8
InternVL2.5-38B
- 30.4 GPT-4o
- 30.3 GPT-4 Turbo
- 30.1
Ovis2-16B
- 29.0
InternVL3-8B
- 27.6
InternVL3-9B
- 27.1 Claude 3 Opus
- 25.9 Qwen2-VL-72B-Instruct
- 25.3 LLaVA-OneVision 72B
- 25.1
Qwen2.5-VL-7B
- 25.0
Qwen2.5-Omni-7B
- 23.4 MiniCPM-V-2_6
- 23.1
InternVL2.5-26B
- 22.8 GPT-4V
- 21.7 MiniCPM-o-2_6
- 21.7
InternVL3-2B
- 21.2
Qwen2.5-VL-3B
- 19.7
InternVL2.5-8B
- 19.2
Gemini 1.5 Pro
- 18.8
InternVL3-1B
- 18.8
Ovis1.6-Gemma2-9B
- 17.9 Aquila-VL-2B
- 17.7
Gemini 1.0 Pro
- 17.0
InternVL-Chat-V1-2-Plus
- 16.3 Qwen2-VL-7B-Instruct
- 15.6 Qwen-VL-Max
- 14.7
InternVL2.5-2B
- 14.5
InternLM-XComposer2-VL-7B
- 14.4
InternVL2.5-1B
- 12.4
Qwen2-VL-2B-Instruct
- 11.9
ShareGPT4V-13B
- 11.1
LLaVA-v1.5-13B
- 10.7
Qwen-VL-Plus
- 10.5
ShareGPT4V-7B
- 8.52 llava-v1.5-7b
- 1
Qwen 3.8 Max 97.7%
- 2
LongCat-2.5 96.5%
- 3
Qwen3.8-Omni-Flash 96.2%
- 4 GPT-5.4 xhigh, with python 96.1%
- 5
Gemini 3.1 Pro thinking high, with python 95.7%
- 6
Qwen3.8 Flash 95.7%
- 7
Qwen 3.8 27B 94.6%
- 8
Seed2.1 Pro 94.5%
- 9 Kimi K3 94.3%
- 10 Kimi K2.6 with python 93.2%
- 11
Inkling 2 Small 92.8%
- 12
Gemini 3 Flash 90.6%
- 13
Qwen3.7-Plus 90.3%
- 14
Seed2.0 Lite (0428) 89.8%
- 15
Inkling 2 Mini 89%
- 16
Seed2.0 Pro 88.8%
- 17
Qwen3.5 397B-A17B 88.6%
- 18
Qwen 3.6 Plus 88%
- 19
Gemini 3 Pro 86.6%
- 20
Seed2.0 Lite (0215) 86.4%
- 21
Qwen3.5-122B-A10B 86.2%
- 22
Qwen3.5-27B 86%
- 23
Gemma 4 31B IT 85.6%
- 24 Kimi K2.5 thinking, with python 85%
- 25 Claude Opus 4.6 max, with python 84.6%
- 26
Qwen3.5-35B-A3B 83.9%
- 27 GPT-5.2 83%
- 28
Gemma 4 26B A4B 82.4%
- 29
Seed1.8 81.3%
- 30
Gemma 4 12B 79.7%
- 31
Qwen3.5-9B 78.9%
- 32
Nemotron 3.5 Super VL 78.2%
- 33
Seed1.6-Thinking 77.2%
- 34 Claude Opus 4.5 77.1%
- 35
EXAONE 4.5 33B 75.2%
- 36
Qwen3-VL-235B-A22B thinking 74.6%
- 37
Qwen3.5-4B 74.6%
- 38
ERNIE 5.0 74.3%
- 39
MiniCPM-V-4.7-35B-A3B 74.2%
- 40
Gemini 2.5 Pro Mar 2025 73.3%
- 41
Qwen3.5-Omni-Plus 73%
- 42 GPT-5 72%
- 43 GPT-5 mini 71.9%
- 44 Claude Sonnet 4.5 71.1%
- 45
STEP3-VL-10B 70.8%
- 46
DiffusionGemma 26B A4B 70.5%
- 47 dots.vlm1 69.6%
- 48
Seed1.5-VL 68.7%
- 49 Claude Opus 4.1 thinking 66%
- 50
Qwen3-VL-235B-A22B-Instruct 66%
- 51 GLM-4.5V 65.6%
- 52
Step 3 64.8%
- 53
InternVL3.5-241B-A28B 63.9%
- 54 GLM-4.6V 63.5%
- 55
Intern-S1 62.5%
- 56
ERNIE-4.5-VL-28B-A3B-Thinking 61.9%
- 57 MiMo-VL-7B-RL 60.4%
- 58 o1 60.3%
- 59 MiMo-VL-7B-RL-2508 59.6%
- 60
Qwen3-VL-8B-Thinking 59.6%
- 61
Gemma 4 E4B 59.5%
- 62 Claude 3.7 Sonnet 58.6%
- 63 o4-mini 58%
- 64 MiMo-VL-7B-SFT 57.9%
- 65 Kimi-VL-A3B-Thinking-2506 56.9%
- 66
Qwen3-Omni-30B-A3B 56.3%
- 67
InternVL3.5-30B-A3B 55.7%
- 68 GLM-4.1V-9B-Thinking 54.4%
- 69 GLM-4.6V-Flash 54%
- 70
InternVL3.5-38B 54%
- 71
Ovis2.5-9B 53.9%
- 72
Nemotron Nano 12B v2 VL 53.6%
- 73
Skywork-R1V3-38B 52.6%
- 74
Gemma 4 E2B 52.4%
- 75
InternVL3.5-8B 52%
- 76
Qianfan-VL-70B 50.3%
- 77
Skywork-R1V2-38B 49.7%
- 78
Doubao-1.5-pro 48.6%
- 79
Gemini 2.0 Pro 48.1%
- 80 GPT-4.5 47.3%
- 81
Nemotron 3.5 Bolt E4B 46.3%
- 82
Gemma 3 27B non-thinking 46%
- 83
Keye-VL-8B-Preview 46%
- 84
InternVL3-78B 43.1%
- 85
Gemini 2.0 Flash Feb 2025 41.3%
- 86
InternVL3.5-4B 40%
- 87 Kimi k1.5 38.6%
- 88
Qwen2.5-VL-72B 38.1%
- 89 Claude 3.5 Sonnet 38%
- 90
Ovis2.5-2B 37.4%
- 91
InternVL3-14B 37%
- 92 Kimi-VL-A3B-Thinking 36.8%
- 93
Phi-4-Reasoning-Vision-15B 36.2%
- 94
QVQ-72B-Preview 35.9%
- 95
InternVL3-38B 34.5%
- 96
Qianfan-VL-8B 32.8%
- 97
InternVL2_5-78B 32.2%
- 98
Ovis2-34B 31.9%
- 99
InternVL2.5-38B 31.8%
- 100 GPT-4o May 2024 30.4%
- 101 GPT-4 Turbo Apr 2024 30.3%
- 102
Ovis2-16B 30.1%
- 103
InternVL3-8B 29%
- 104
InternVL3-9B 27.6%
- 105 Claude 3 Opus 27.1%
- 106 Qwen2-VL-72B-Instruct 25.9%
- 107 LLaVA-OneVision 72B 25.3%
- 108
Qwen2.5-VL-7B 25.1%
- 109
Qwen2.5-Omni-7B 25%
- 110 MiniCPM-V-2_6 23.4%
- 111
InternVL2.5-26B 23.1%
- 112 GPT-4V 22.8%
- 113 MiniCPM-o-2_6 21.7%
- 114
InternVL3-2B 21.7%
- 115
Qwen2.5-VL-3B 21.2%
- 116
InternVL2.5-8B 19.7%
- 117
Gemini 1.5 Pro May 2024 19.2%
- 118
InternVL3-1B 18.8%
- 119
Ovis1.6-Gemma2-9B 18.8%
- 120 Aquila-VL-2B 17.9%
- 121
Gemini 1.0 Pro 17.7%
- 122
InternVL-Chat-V1-2-Plus 17%
- 123 Qwen2-VL-7B-Instruct 16.3%
- 124 Qwen-VL-Max 15.6%
- 125
InternVL2.5-2B 14.7%
- 126
InternLM-XComposer2-VL-7B 14.5%
- 127
InternVL2.5-1B 14.4%
- 128
Qwen2-VL-2B-Instruct 12.4%
- 129
ShareGPT4V-13B 11.9%
- 130
LLaVA-v1.5-13B 11.1%
- 131
Qwen-VL-Plus 10.7%
- 132
ShareGPT4V-7B 10.5%
- 133 llava-v1.5-7b 8.52%
Sources: Labs’ model cards and official MATH-Vision (MathVision) leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 97.7% | Aug 2026 | Lab model card | |
| 2 | | 96.5% | Nov 2026 | Lab model card | |
| 3 | | 96.2% | Sep 2026 | Lab model card | |
| 4 | | 95.7% | Aug 2026 | Lab model card | |
| 5 | | 94.6% | Aug 2026 | Lab model card | |
| 6 | | 94.5% | Jun 2026 | Lab model card | |
| 7 | Kimi K3 | 94.3% | Jul 2026 | Lab model card | |
| 8 | | 92.8% | Nov 2026 | Lab model card | |
| 9 | | 90.3% | Jun 2026 | Lab model card | |
| 10 | | 89.8% | Apr 2026 | Lab model card | |
| 11 | | 89% | Nov 2026 | Lab model card | |
| 12 | | 88.8% | Feb 2026 | Lab model card | |
| 13 | | 88.6% | Feb 2026 | Lab model card | |
| 14 | | 88% | Apr 2026 | Lab model card | |
| 15 | Kimi K2.6 | 87.4% | Apr 2026 | Lab model card | |
| 16 | | 86.2% | Feb 2026 | Lab model card | |
| 17 | | 86% | Feb 2026 | Lab model card | |
| 18 | | 85.6% | Apr 2026 | Lab model card | |
| 19 | Kimi K2.5 | 84.2% | Jan 2026 | Lab model card | |
| 20 | | 83.9% | Feb 2026 | Lab model card | |
| 21 | | 82.4% | Apr 2026 | Lab model card | |
| 22 | | 81.3% | Dec 2025 | Lab model card | |
| 23 | | 79.7% | Jun 2026 | Lab model card | |
| 24 | | 78.9% | Mar 2026 | Lab model card | |
| 25 | | 78.2% | Oct 2026 | Lab model card | |
| 26 | | 75.2% | Apr 2026 | Lab model card | |
| 27 | | 74.6% | Sep 2025 | Lab model card | |
| 28 | | 74.6% | Mar 2026 | Lab model card | |
| 29 | | 74.3% | Feb 2026 | Lab model card | |
| 30 | | 74.2% | Oct 2026 | Lab model card | |
| 31 | | 73% | Mar 2026 | Lab model card | |
| 32 | | 70.8% | Jan 2026 | Lab model card | |
| 33 | | 70.5% | Jun 2026 | Lab model card | |
| 34 | | 70.3% | Jun 2025 | Lab model card | |
| 35 | | 68.7% | May 2025 | Lab model card | |
| 36 | GLM-4.5V | 65.6% | Aug 2025 | Lab model card | |
| 37 | | 64.8% | Jul 2025 | Lab model card | |
| 38 | | 63.9% | Aug 2025 | Lab model card | |
| 39 | GLM-4.6V | 63.5% | Jan 2026 | Lab model card | |
| 40 | | 62.5% | Jul 2025 | Lab model card | |
| 41 | | 61.9% | Nov 2025 | Lab model card | |
| 42 | | 59.5% | Apr 2026 | Lab model card | |
| 43 | Kimi-VL-A3B-Thinking-2506 | 56.9% | Jun 2025 | Lab model card | |
| 44 | | 56.3% | Sep 2025 | Lab model card | |
| 45 | GLM-4.1V-9B-Thinking | 54.4% | Jul 2025 | Lab model card | |
| 46 | | 53.6% | Nov 2025 | Lab model card | |
| 47 | | 52.4% | Apr 2026 | Lab model card | |
| 48 | | 46.3% | Oct 2026 | Lab model card | |
| 49 | Kimi k1.5 | 38.6% | Jan 2025 | Lab model card | |
| 50 | | 38.1% | Jan 2025 | Lab model card | |
| 51 | Kimi-VL-A3B-Thinking | 36.8% | Apr 2025 | Lab model card | |
| 52 | | 36.2% | Mar 2026 | Lab model card | |
| 53 | | 25% | Mar 2025 | Lab model card |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5.4 | 96.1% | xhigh, with python | Mar 2026 | Official leaderboard |
| 2 | 95.7% | thinking high, with python | Feb 2026 | Official leaderboard | |
| 3 | Kimi K2.6 | 93.2% | with python | Apr 2026 | Official leaderboard |
| 4 | 90.6% | Dec 2025 | Official leaderboard | ||
| 5 | 89.8% | Apr 2026 | Official leaderboard | ||
| 6 | 88.8% | Feb 2026 | Official leaderboard | ||
| 7 | 88.6% | Feb 2026 | Official leaderboard | ||
| 8 | 86.6% | Nov 2025 | Official leaderboard | ||
| 9 | 86.4% | Feb 2026 | Official leaderboard | ||
| 10 | 86.2% | Feb 2026 | Official leaderboard | ||
| 11 | 86% | Feb 2026 | Official leaderboard | ||
| 12 | 85.6% | Apr 2026 | Official leaderboard | ||
| 13 | Kimi K2.5 | 85% | thinking, with python | Jan 2026 | Official leaderboard |
| 14 | Claude Opus 4.6 | 84.6% | max, with python | Feb 2026 | Official leaderboard |
| 15 | 83.9% | Feb 2026 | Official leaderboard | ||
| 16 | GPT-5.2 | 83% | Dec 2025 | Official leaderboard | |
| 17 | 82.4% | Apr 2026 | Official leaderboard | ||
| 18 | 81.3% | Dec 2025 | Official leaderboard | ||
| 19 | 77.2% | Jun 2025 | Official leaderboard | ||
| 20 | Claude Opus 4.5 | 77.1% | Nov 2025 | Official leaderboard | |
| 21 | 74.6% | thinking | Sep 2025 | Official leaderboard | |
| 22 | 73.3% | Mar 2025 | Official leaderboard | ||
| 23 | GPT-5 | 72% | Aug 2025 | Official leaderboard | |
| 24 | GPT-5 mini | 71.9% | Aug 2025 | Official leaderboard | |
| 25 | Claude Sonnet 4.5 | 71.1% | Sep 2025 | Official leaderboard | |
| 26 | dots.vlm1 | 69.6% | Aug 2025 | Official leaderboard | |
| 27 | 68.7% | May 2025 | Official leaderboard | ||
| 28 | Claude Opus 4.1 | 66% | thinking | Aug 2025 | Official leaderboard |
| 29 | 66% | Sep 2025 | Official leaderboard | ||
| 30 | GLM-4.5V | 65.6% | Aug 2025 | Official leaderboard | |
| 31 | 64.8% | Jul 2025 | Official leaderboard | ||
| 32 | 63.9% | Aug 2025 | Official leaderboard | ||
| 33 | GLM-4.6V | 63.5% | Dec 2025 | Official leaderboard | |
| 34 | MiMo-VL-7B-RL | 60.4% | May 2025 | Official leaderboard | |
| 35 | o1 | 60.3% | Dec 2024 | Official leaderboard | |
| 36 | MiMo-VL-7B-RL-2508 | 59.6% | Aug 2025 | Official leaderboard | |
| 37 | 59.6% | Oct 2025 | Official leaderboard | ||
| 38 | 59.5% | Mar 2026 | Official leaderboard | ||
| 39 | Claude 3.7 Sonnet | 58.6% | Feb 2025 | Official leaderboard | |
| 40 | o4-mini | 58% | Apr 2025 | Official leaderboard | |
| 41 | MiMo-VL-7B-SFT | 57.9% | May 2025 | Official leaderboard | |
| 42 | Kimi-VL-A3B-Thinking-2506 | 56.9% | Jun 2025 | Official leaderboard | |
| 43 | 55.7% | Aug 2025 | Official leaderboard | ||
| 44 | GLM-4.1V-9B-Thinking | 54.4% | Jul 2025 | Official leaderboard | |
| 45 | GLM-4.6V-Flash | 54% | Dec 2025 | Official leaderboard | |
| 46 | 54% | Aug 2025 | Official leaderboard | ||
| 47 | 53.9% | Aug 2025 | Official leaderboard | ||
| 48 | 52.6% | Jul 2025 | Official leaderboard | ||
| 49 | 52.4% | Mar 2026 | Official leaderboard | ||
| 50 | 52% | Aug 2025 | Official leaderboard | ||
| 51 | 50.3% | Sep 2025 | Official leaderboard | ||
| 52 | 49.7% | Apr 2025 | Official leaderboard | ||
| 53 | 48.6% | Jan 2025 | Official leaderboard | ||
| 54 | 48.1% | Feb 2025 | Official leaderboard | ||
| 55 | GPT-4.5 | 47.3% | Feb 2025 | Official leaderboard | |
| 56 | 46% | non-thinking | Mar 2025 | Official leaderboard | |
| 57 | 46% | Jun 2025 | Official leaderboard | ||
| 58 | 43.1% | Apr 2025 | Official leaderboard | ||
| 59 | 41.3% | Feb 2025 | Official leaderboard | ||
| 60 | 40% | Aug 2025 | Official leaderboard | ||
| 61 | Kimi k1.5 | 38.6% | Jan 2025 | Official leaderboard | |
| 62 | 38.1% | Jan 2025 | Official leaderboard | ||
| 63 | Claude 3.5 Sonnet | 38% | Jun 2024 | Official leaderboard | |
| 64 | 37.4% | Aug 2025 | Official leaderboard | ||
| 65 | 37% | Apr 2025 | Official leaderboard | ||
| 66 | Kimi-VL-A3B-Thinking | 36.8% | Apr 2025 | Official leaderboard | |
| 67 | 35.9% | Dec 2024 | Official leaderboard | ||
| 68 | 34.5% | Apr 2025 | Official leaderboard | ||
| 69 | 32.8% | Sep 2025 | Official leaderboard | ||
| 70 | 32.2% | Dec 2024 | Official leaderboard | ||
| 71 | 31.9% | Feb 2025 | Official leaderboard | ||
| 72 | 31.8% | Dec 2024 | Official leaderboard | ||
| 73 | GPT-4o May 2024 | 30.4% | May 2024 | Official leaderboard | |
| 74 | GPT-4 Turbo Apr 2024 | 30.3% | Apr 2024 | Official leaderboard | |
| 75 | 30.1% | Feb 2025 | Official leaderboard | ||
| 76 | 29% | Apr 2025 | Official leaderboard | ||
| 77 | 27.6% | Apr 2025 | Official leaderboard | ||
| 78 | Claude 3 Opus | 27.1% | Feb 2024 | Official leaderboard | |
| 79 | Qwen2-VL-72B-Instruct | 25.9% | Aug 2024 | Official leaderboard | |
| 80 | LLaVA-OneVision 72B | 25.3% | Aug 2024 | Official leaderboard | |
| 81 | 25.1% | Jan 2025 | Official leaderboard | ||
| 82 | MiniCPM-V-2_6 | 23.4% | Aug 2024 | Official leaderboard | |
| 83 | 23.1% | Dec 2024 | Official leaderboard | ||
| 84 | GPT-4V | 22.8% | Nov 2023 | Official leaderboard | |
| 85 | MiniCPM-o-2_6 | 21.7% | Jan 2025 | Official leaderboard | |
| 86 | 21.7% | Apr 2025 | Official leaderboard | ||
| 87 | 21.2% | Jan 2025 | Official leaderboard | ||
| 88 | 19.7% | Dec 2024 | Official leaderboard | ||
| 89 | 19.2% | May 2024 | Official leaderboard | ||
| 90 | 18.8% | Apr 2025 | Official leaderboard | ||
| 91 | 18.8% | Sep 2024 | Official leaderboard | ||
| 92 | Aquila-VL-2B | 17.9% | Oct 2024 | Official leaderboard | |
| 93 | 17.7% | Dec 2023 | Official leaderboard | ||
| 94 | 17% | Feb 2024 | Official leaderboard | ||
| 95 | Qwen2-VL-7B-Instruct | 16.3% | Aug 2024 | Official leaderboard | |
| 96 | Qwen-VL-Max | 15.6% | Jan 2024 | Official leaderboard | |
| 97 | 14.7% | Dec 2024 | Official leaderboard | ||
| 98 | 14.5% | Jan 2024 | Official leaderboard | ||
| 99 | 14.4% | Dec 2024 | Official leaderboard | ||
| 100 | 12.4% | Aug 2024 | Official leaderboard | ||
| 101 | 11.9% | Dec 2023 | Official leaderboard | ||
| 102 | 11.1% | Oct 2023 | Official leaderboard | ||
| 103 | 10.7% | Jan 2024 | Official leaderboard | ||
| 104 | 10.5% | Nov 2023 | Official leaderboard | ||
| 105 | llava-v1.5-7b | 8.52% | Oct 2023 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | 97.7% | Aug 2026 | Lab model card | ||
| 2 | 96.5% | Nov 2026 | Lab model card | ||
| 3 | 96.2% | Sep 2026 | Lab model card | ||
| 4 | GPT-5.4 | 96.1% | xhigh, with python | Mar 2026 | Official leaderboard |
| 5 | 95.7% | thinking high, with python | Feb 2026 | Official leaderboard | |
| 6 | 95.7% | Aug 2026 | Lab model card | ||
| 7 | 94.6% | Aug 2026 | Lab model card | ||
| 8 | 94.5% | Jun 2026 | Lab model card | ||
| 9 | Kimi K3 | 94.3% | Jul 2026 | Lab model card | |
| 10 | Kimi K2.6 | 93.2% | with python | Apr 2026 | Official leaderboard |
| 11 | 92.8% | Nov 2026 | Lab model card | ||
| 12 | 90.6% | Dec 2025 | Official leaderboard | ||
| 13 | 90.3% | Jun 2026 | Lab model card | ||
| 14 | 89.8% | Apr 2026 | Official leaderboard | ||
| 15 | 89% | Nov 2026 | Lab model card | ||
| 16 | 88.8% | Feb 2026 | Official leaderboard | ||
| 17 | 88.6% | Feb 2026 | Official leaderboard | ||
| 18 | 88% | Apr 2026 | Lab model card | ||
| 19 | 86.6% | Nov 2025 | Official leaderboard | ||
| 20 | 86.4% | Feb 2026 | Official leaderboard | ||
| 21 | 86.2% | Feb 2026 | Official leaderboard | ||
| 22 | 86% | Feb 2026 | Official leaderboard | ||
| 23 | 85.6% | Apr 2026 | Official leaderboard | ||
| 24 | Kimi K2.5 | 85% | thinking, with python | Jan 2026 | Official leaderboard |
| 25 | Claude Opus 4.6 | 84.6% | max, with python | Feb 2026 | Official leaderboard |
| 26 | 83.9% | Feb 2026 | Official leaderboard | ||
| 27 | GPT-5.2 | 83% | Dec 2025 | Official leaderboard | |
| 28 | 82.4% | Apr 2026 | Official leaderboard | ||
| 29 | 81.3% | Dec 2025 | Official leaderboard | ||
| 30 | 79.7% | Jun 2026 | Lab model card | ||
| 31 | 78.9% | Mar 2026 | Lab model card | ||
| 32 | 78.2% | Oct 2026 | Lab model card | ||
| 33 | 77.2% | Jun 2025 | Official leaderboard | ||
| 34 | Claude Opus 4.5 | 77.1% | Nov 2025 | Official leaderboard | |
| 35 | 75.2% | Apr 2026 | Lab model card | ||
| 36 | 74.6% | thinking | Sep 2025 | Official leaderboard | |
| 37 | 74.6% | Mar 2026 | Lab model card | ||
| 38 | 74.3% | Feb 2026 | Lab model card | ||
| 39 | 74.2% | Oct 2026 | Lab model card | ||
| 40 | 73.3% | Mar 2025 | Official leaderboard | ||
| 41 | 73% | Mar 2026 | Lab model card | ||
| 42 | GPT-5 | 72% | Aug 2025 | Official leaderboard | |
| 43 | GPT-5 mini | 71.9% | Aug 2025 | Official leaderboard | |
| 44 | Claude Sonnet 4.5 | 71.1% | Sep 2025 | Official leaderboard | |
| 45 | 70.8% | Jan 2026 | Lab model card | ||
| 46 | 70.5% | Jun 2026 | Lab model card | ||
| 47 | dots.vlm1 | 69.6% | Aug 2025 | Official leaderboard | |
| 48 | 68.7% | May 2025 | Official leaderboard | ||
| 49 | Claude Opus 4.1 | 66% | thinking | Aug 2025 | Official leaderboard |
| 50 | 66% | Sep 2025 | Official leaderboard | ||
| 51 | GLM-4.5V | 65.6% | Aug 2025 | Official leaderboard | |
| 52 | 64.8% | Jul 2025 | Official leaderboard | ||
| 53 | 63.9% | Aug 2025 | Official leaderboard | ||
| 54 | GLM-4.6V | 63.5% | Dec 2025 | Official leaderboard | |
| 55 | 62.5% | Jul 2025 | Lab model card | ||
| 56 | 61.9% | Nov 2025 | Lab model card | ||
| 57 | MiMo-VL-7B-RL | 60.4% | May 2025 | Official leaderboard | |
| 58 | o1 | 60.3% | Dec 2024 | Official leaderboard | |
| 59 | MiMo-VL-7B-RL-2508 | 59.6% | Aug 2025 | Official leaderboard | |
| 60 | 59.6% | Oct 2025 | Official leaderboard | ||
| 61 | 59.5% | Mar 2026 | Official leaderboard | ||
| 62 | Claude 3.7 Sonnet | 58.6% | Feb 2025 | Official leaderboard | |
| 63 | o4-mini | 58% | Apr 2025 | Official leaderboard | |
| 64 | MiMo-VL-7B-SFT | 57.9% | May 2025 | Official leaderboard | |
| 65 | Kimi-VL-A3B-Thinking-2506 | 56.9% | Jun 2025 | Official leaderboard | |
| 66 | 56.3% | Sep 2025 | Lab model card | ||
| 67 | 55.7% | Aug 2025 | Official leaderboard | ||
| 68 | GLM-4.1V-9B-Thinking | 54.4% | Jul 2025 | Official leaderboard | |
| 69 | GLM-4.6V-Flash | 54% | Dec 2025 | Official leaderboard | |
| 70 | 54% | Aug 2025 | Official leaderboard | ||
| 71 | 53.9% | Aug 2025 | Official leaderboard | ||
| 72 | 53.6% | Nov 2025 | Lab model card | ||
| 73 | 52.6% | Jul 2025 | Official leaderboard | ||
| 74 | 52.4% | Mar 2026 | Official leaderboard | ||
| 75 | 52% | Aug 2025 | Official leaderboard | ||
| 76 | 50.3% | Sep 2025 | Official leaderboard | ||
| 77 | 49.7% | Apr 2025 | Official leaderboard | ||
| 78 | 48.6% | Jan 2025 | Official leaderboard | ||
| 79 | 48.1% | Feb 2025 | Official leaderboard | ||
| 80 | GPT-4.5 | 47.3% | Feb 2025 | Official leaderboard | |
| 81 | 46.3% | Oct 2026 | Lab model card | ||
| 82 | 46% | non-thinking | Mar 2025 | Official leaderboard | |
| 83 | 46% | Jun 2025 | Official leaderboard | ||
| 84 | 43.1% | Apr 2025 | Official leaderboard | ||
| 85 | 41.3% | Feb 2025 | Official leaderboard | ||
| 86 | 40% | Aug 2025 | Official leaderboard | ||
| 87 | Kimi k1.5 | 38.6% | Jan 2025 | Official leaderboard | |
| 88 | 38.1% | Jan 2025 | Official leaderboard | ||
| 89 | Claude 3.5 Sonnet | 38% | Jun 2024 | Official leaderboard | |
| 90 | 37.4% | Aug 2025 | Official leaderboard | ||
| 91 | 37% | Apr 2025 | Official leaderboard | ||
| 92 | Kimi-VL-A3B-Thinking | 36.8% | Apr 2025 | Official leaderboard | |
| 93 | 36.2% | Mar 2026 | Lab model card | ||
| 94 | 35.9% | Dec 2024 | Official leaderboard | ||
| 95 | 34.5% | Apr 2025 | Official leaderboard | ||
| 96 | 32.8% | Sep 2025 | Official leaderboard | ||
| 97 | 32.2% | Dec 2024 | Official leaderboard | ||
| 98 | 31.9% | Feb 2025 | Official leaderboard | ||
| 99 | 31.8% | Dec 2024 | Official leaderboard | ||
| 100 | GPT-4o May 2024 | 30.4% | May 2024 | Official leaderboard | |
| 101 | GPT-4 Turbo Apr 2024 | 30.3% | Apr 2024 | Official leaderboard | |
| 102 | 30.1% | Feb 2025 | Official leaderboard | ||
| 103 | 29% | Apr 2025 | Official leaderboard | ||
| 104 | 27.6% | Apr 2025 | Official leaderboard | ||
| 105 | Claude 3 Opus | 27.1% | Feb 2024 | Official leaderboard | |
| 106 | Qwen2-VL-72B-Instruct | 25.9% | Aug 2024 | Official leaderboard | |
| 107 | LLaVA-OneVision 72B | 25.3% | Aug 2024 | Official leaderboard | |
| 108 | 25.1% | Jan 2025 | Official leaderboard | ||
| 109 | 25% | Mar 2025 | Lab model card | ||
| 110 | MiniCPM-V-2_6 | 23.4% | Aug 2024 | Official leaderboard | |
| 111 | 23.1% | Dec 2024 | Official leaderboard | ||
| 112 | GPT-4V | 22.8% | Nov 2023 | Official leaderboard | |
| 113 | MiniCPM-o-2_6 | 21.7% | Jan 2025 | Official leaderboard | |
| 114 | 21.7% | Apr 2025 | Official leaderboard | ||
| 115 | 21.2% | Jan 2025 | Official leaderboard | ||
| 116 | 19.7% | Dec 2024 | Official leaderboard | ||
| 117 | 19.2% | May 2024 | Official leaderboard | ||
| 118 | 18.8% | Apr 2025 | Official leaderboard | ||
| 119 | 18.8% | Sep 2024 | Official leaderboard | ||
| 120 | Aquila-VL-2B | 17.9% | Oct 2024 | Official leaderboard | |
| 121 | 17.7% | Dec 2023 | Official leaderboard | ||
| 122 | 17% | Feb 2024 | Official leaderboard | ||
| 123 | Qwen2-VL-7B-Instruct | 16.3% | Aug 2024 | Official leaderboard | |
| 124 | Qwen-VL-Max | 15.6% | Jan 2024 | Official leaderboard | |
| 125 | 14.7% | Dec 2024 | Official leaderboard | ||
| 126 | 14.5% | Jan 2024 | Official leaderboard | ||
| 127 | 14.4% | Dec 2024 | Official leaderboard | ||
| 128 | 12.4% | Aug 2024 | Official leaderboard | ||
| 129 | 11.9% | Dec 2023 | Official leaderboard | ||
| 130 | 11.1% | Oct 2023 | Official leaderboard | ||
| 131 | 10.7% | Jan 2024 | Official leaderboard | ||
| 132 | 10.5% | Nov 2023 | Official leaderboard | ||
| 133 | llava-v1.5-7b | 8.52% | Oct 2023 | Official leaderboard |
The benchmark's own leaderboard has 153 entries, 105 of them models we track. Switch the scores above to rank them together.
Rankings by part
We read 16 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Angle, Area and Algebra. Skywork-R1V2-38B leads 8 of them, and 3 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Angle
16 models we track have a score here.
- 1
Skywork-R1V2-38B ↑51 63.2%
- 2
Doubao-1.5-pro ↑51 62.4%
- 3 GPT-4 Turbo Apr 2024 ↑71 30.6%
- 4
InternVL-Chat-V1-2-Plus ↑90 22.5%
- 5 GPT-4V ↑79 22%
- 6
Ovis1.6-Gemma2-9B ↑85 20.8%
- 7
InternLM-XComposer2-VL-7B ↑91 19.7%
- 8
Gemini 1.0 Pro ↑85 19.1%
- 9 GPT-4o May 2024 ↑64 17.3%
- 10
ShareGPT4V-7B ↑94 16.2%
Area
16 models we track have a score here.
- 1
Skywork-R1V2-38B ↑51 73.7%
- 2
Doubao-1.5-pro ↑51 55.4%
- 3 GPT-4o May 2024 ↑70 29.8%
- 4 GPT-4 Turbo Apr 2024 ↑70 29%
- 5 GPT-4V ↑79 22.2%
- 6
Gemini 1.0 Pro ↑87 19%
- 7
Ovis1.6-Gemma2-9B ↑84 18%
- 8
InternVL-Chat-V1-2-Plus ↑86 16.4%
- 9
InternLM-XComposer2-VL-7B ↑89 15.6%
- 10
Gemini 1.5 Pro May 2024 ↑79 14.4%
Algebra
16 models we track have a score here.
- 1
Doubao-1.5-pro ↑52 55.1%
- 2
Skywork-R1V2-38B ↑50 52.6%
- 3 GPT-4o May 2024 ↑70 42%
- 4 GPT-4 Turbo Apr 2024 ↑70 37.7%
- 5 GPT-4V ↑79 27.3%
- 6
Gemini 1.5 Pro May 2024 ↑83 20.3%
- 7
Gemini 1.0 Pro ↑86 15.1%
- 8
Ovis1.6-Gemma2-9B ↑83 13.3%
- 9
Qwen-VL-Plus ↑94 11.3%
- 10
InternVL-Chat-V1-2-Plus ↑84 11.3%
Analytic geometry
16 models we track have a score here.
- 1
Doubao-1.5-pro ↑52 52.4%
- 2
Skywork-R1V2-38B ↑50 47.4%
- 3 GPT-4o May 2024 ↑70 39.3%
- 4
Gemini 1.5 Pro May 2024 ↑85 35.7%
- 5 GPT-4 Turbo Apr 2024 ↑69 33.3%
- 6 GPT-4V ↑78 32.1%
- 7
InternVL-Chat-V1-2-Plus ↑87 25%
- 8 Qwen-VL-Max ↑88 19.1%
- 9
Qwen-VL-Plus ↑94 17.9%
- 10
ShareGPT4V-13B ↑91 15.5%
Arithmetic
16 models we track have a score here.
- 1
Skywork-R1V2-38B ↑51 73.7%
- 2
Doubao-1.5-pro ↑51 63.6%
- 3 GPT-4o May 2024 ↑70 49.3%
- 4 GPT-4 Turbo Apr 2024 ↑70 46.4%
- 5 GPT-4V ↑79 35.7%
- 6
Gemini 1.5 Pro May 2024 ↑83 34.3%
- 7
Ovis1.6-Gemma2-9B ↑84 22.1%
- 8
Gemini 1.0 Pro ↑85 20.7%
- 9 Qwen-VL-Max ↑87 20%
- 10
ShareGPT4V-13B ↑91 16.4%
Combinatorial geometry
16 models we track have a score here.
- 1
Skywork-R1V2-38B ↑51 42.1%
- 2
Doubao-1.5-pro ↑51 34.7%
- 3 GPT-4o May 2024 ↑70 28.9%
- 4 GPT-4 Turbo Apr 2024 ↑70 25%
- 5 GPT-4V ↑79 21.1%
- 6
Gemini 1.0 Pro ↑87 20.1%
- 7
Gemini 1.5 Pro May 2024 ↑82 19.8%
- 8
Ovis1.6-Gemma2-9B ↑83 17.9%
- 9 Qwen-VL-Max ↑87 16.9%
- 10
InternVL-Chat-V1-2-Plus ↑84 16.9%
Combinatorics
16 models we track have a score here.
- 1
Skywork-R1V2-38B ↑51 52.6%
- 2
Doubao-1.5-pro ↑51 36.9%
- 3 GPT-4 Turbo Apr 2024 ↑71 28.6%
- 4 GPT-4o May 2024 ↑69 25.6%
- 5 GPT-4V ↑79 16.7%
- 6
Gemini 1.5 Pro May 2024 ↑83 15.5%
- 7 Qwen-VL-Max ↑89 12.5%
- 8
Gemini 1.0 Pro ↑85 11.9%
- 9
InternLM-XComposer2-VL-7B ↑89 11.3%
- 10
Ovis1.6-Gemma2-9B ↑81 11.3%
Counting
16 models we track have a score here.
- 1
Doubao-1.5-pro ↑52 43.3%
- 2
Skywork-R1V2-38B ↑50 36.8%
- 3 GPT-4 Turbo Apr 2024 ↑71 25.3%
- 4 GPT-4o May 2024 ↑69 22.4%
- 5
Ovis1.6-Gemma2-9B ↑86 22.4%
- 6
Gemini 1.5 Pro May 2024 ↑83 20.9%
- 7 Qwen-VL-Max ↑89 17.9%
- 8 GPT-4V ↑76 13.4%
- 9
InternVL-Chat-V1-2-Plus ↑85 11.9%
- 10 llava-v1.5-7b ↑95 10.5%
Descriptive geometry
16 models we track have a score here.
- 1
Gemini 1.5 Pro May 2024 ↑88 26%
- 2
Doubao-1.5-pro ↑51 25%
- 3 GPT-4o May 2024 ↑70 24%
- 4
Ovis1.6-Gemma2-9B ↑87 23.1%
- 5 GPT-4V ↑79 22.1%
- 6
Gemini 1.0 Pro ↑87 20.2%
- 7 Qwen-VL-Max ↑89 16.4%
- 8
InternVL-Chat-V1-2-Plus ↑86 16.4%
- 9
Skywork-R1V2-38B ↑43 15.8%
- 10
Qwen-VL-Plus ↑93 15.4%
Graph theory
16 models we track have a score here.
- 1
Skywork-R1V2-38B ↑51 57.9%
- 2 GPT-4 Turbo Apr 2024 ↑72 27.8%
- 3
Doubao-1.5-pro ↑50 27.8%
- 4
Gemini 1.5 Pro May 2024 ↑85 26.7%
- 5 GPT-4o May 2024 ↑68 23.3%
- 6
InternLM-XComposer2-VL-7B ↑92 22.2%
- 7
Gemini 1.0 Pro ↑86 21.1%
- 8
Ovis1.6-Gemma2-9B ↑83 20%
- 9
InternVL-Chat-V1-2-Plus ↑85 15.6%
- 10 GPT-4V ↑74 14.4%
Logic
16 models we track have a score here.
- 1
Skywork-R1V2-38B ↑51 73.7%
- 2
Doubao-1.5-pro ↑51 37.8%
- 3 GPT-4 Turbo Apr 2024 ↑71 31.9%
- 4 GPT-4o May 2024 ↑69 29.4%
- 5
Gemini 1.5 Pro May 2024 ↑84 22.7%
- 6 Qwen-VL-Max ↑90 21%
- 7
Ovis1.6-Gemma2-9B ↑84 20.2%
- 8
InternLM-XComposer2-VL-7B ↑90 19.3%
- 9
InternVL-Chat-V1-2-Plus ↑85 19.3%
- 10 GPT-4V ↑74 16.8%
Length
16 models we track have a score here.
- 1
Doubao-1.5-pro ↑52 59.7%
- 2
Skywork-R1V2-38B ↑50 57.9%
- 3 GPT-4 Turbo Apr 2024 ↑71 31.9%
- 4 GPT-4o May 2024 ↑69 30.1%
- 5
Ovis1.6-Gemma2-9B ↑86 24.7%
- 6
InternVL-Chat-V1-2-Plus ↑88 22.5%
- 7 GPT-4V ↑77 20.9%
- 8
Gemini 1.0 Pro ↑85 20%
- 9 Qwen-VL-Max ↑87 19.8%
- 10
ShareGPT4V-13B ↑91 17.4%
Solid geometry
16 models we track have a score here.
- 1
Skywork-R1V2-38B ↑51 47.4%
- 2
Doubao-1.5-pro ↑51 43.9%
- 3 GPT-4o May 2024 ↑70 29.1%
- 4 GPT-4 Turbo Apr 2024 ↑70 28.7%
- 5 GPT-4V ↑79 23.8%
- 6
Gemini 1.5 Pro May 2024 ↑83 18.9%
- 7
Ovis1.6-Gemma2-9B ↑84 15.6%
- 8
Gemini 1.0 Pro ↑85 14.3%
- 9
Qwen-VL-Plus ↑94 14.3%
- 10
InternVL-Chat-V1-2-Plus ↑84 14.3%
Statistics
16 models we track have a score here.
- 1
Doubao-1.5-pro ↑52 55.2%
- 2
Skywork-R1V2-38B ↑50 47.4%
- 3 GPT-4o May 2024 ↑70 44.8%
- 4 GPT-4 Turbo Apr 2024 ↑70 37.9%
- 5 GPT-4V ↑79 24.1%
- 6 Qwen-VL-Max ↑90 20.7%
- 7
Ovis1.6-Gemma2-9B ↑84 20.7%
- 8
InternVL-Chat-V1-2-Plus ↑86 17.2%
- 9
ShareGPT4V-7B ↑95 15.5%
- 10
InternLM-XComposer2-VL-7B ↑88 15.5%
Topology
16 models we track have a score here.
- 1 GPT-4o May 2024 ↑72 34.8%
- 2
InternLM-XComposer2-VL-7B ↑96 26.1%
- 3
Gemini 1.5 Pro May 2024 ↑86 26.1%
- 4
Doubao-1.5-pro ↑49 26.1%
- 5 GPT-4V ↑79 21.7%
- 6
Skywork-R1V2-38B ↑46 21.1%
- 7
ShareGPT4V-7B ↑97 17.4%
- 8
Gemini 1.0 Pro ↑85 17.4%
- 9 GPT-4 Turbo Apr 2024 ↑65 17.4%
- 10
Ovis1.6-Gemma2-9B ↑81 17.4%
Transformation geometry
16 models we track have a score here.
- 1
Doubao-1.5-pro ↑52 37.5%
- 2
Skywork-R1V2-38B ↑50 31.6%
- 3 GPT-4V ↑81 25.6%
- 4 GPT-4 Turbo Apr 2024 ↑70 23.2%
- 5
Gemini 1.0 Pro ↑88 20.8%
- 6
InternVL-Chat-V1-2-Plus ↑88 20.8%
- 7
Ovis1.6-Gemma2-9B ↑84 20.8%
- 8 GPT-4o May 2024 ↑65 17.9%
- 9 Qwen-VL-Max ↑87 17.3%
- 10
Gemini 1.5 Pro May 2024 ↑79 17.3%
More effort adds a median 6 points
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: Official MATH-Vision (MathVision) leaderboard
11 models have held number one in labs’ own numbers since its release in February 2024 12 models have held number one across all independent runs since its release in February 2024 13 models have held number one with claims included since its release in February 2024
Lab claims count from the day the card came out, independent runs from the model’s release. We draw one row each for the benchmark’s own leaderboard and what labs claim in their cards.
Sources: Official MATH-Vision (MathVision) leaderboard and labs’ model cards
Gemini 3 Pro on the official leaderboard passed 85% in November 2025
That was 21 months after its release. By the same age, 34% of the benchmarks from 2024 had passed 85%. All the scores we have, one dot each, so a model tested at several settings appears more than once.
Sources: Official MATH-Vision (MathVision) leaderboard and labs’ model cards
8 labs put it in a launch post in the last six months
Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.
| Date | Launch post | Score | As the post writes it |
|---|---|---|---|
| Nov 5, 2026 | 92.8 | MathVision, 92.8% | |
| Nov 2, 2026 | 96.5 | MathVision (w/ tools) | |
| Oct 15, 2026 | 78.2 | MathVision | |
| Oct 7, 2026 | 74.2 | MathVision | |
| Sep 18, 2026 | 96.2 | MathVision Without CI 91.8 With CI 96.2 | |
| Aug 26, 2026 | 95.7 | Visual math problem solving MathVision Without CI 90.6 With CI 95.7 | |
| Aug 14, 2026 | 94.6 | Visual math problem solving MathVision Without CI 90.0 With CI 94.6 | |
| Aug 3, 2026 | 97.7 | MathVision (without CI / with CI) 95.2 / 97.7 | |
| Jul 16, 2026 | Kimi K3 | 94.3 | MathVision (without python) |
| Jun 23, 2026 | 94.5 | MathVision (w. Tool) 92.6 (94.5) | |
| Jun 10, 2026 | 70.5 | MATH-Vision | |
| Jun 3, 2026 | 79.7 | MATH-Vision | |
| Jun 1, 2026 | 90.3 | MathVision | |
| Apr 28, 2026 | 89.8 | MathVision | |
| Apr 20, 2026 | Kimi K2.6 | 87.4 | MathVision |
| Apr 9, 2026 | 75.2 | MathVision | |
| Apr 2, 2026 | 59.5 | MATH-Vision | |
| Apr 2, 2026 | 52.4 | MATH-Vision | |
| Apr 2, 2026 | 82.4 | MATH-Vision | |
| Apr 2, 2026 | 88 | MathVision | |
| Apr 2, 2026 | 85.6 | MATH-Vision | |
| Mar 30, 2026 | 73 | MathVision | |
| Mar 4, 2026 | 36.2 | MathVision_MINI (default mode) | |
| Mar 2, 2026 | 74.6 | MathVision | |
| Mar 2, 2026 | 78.9 | MathVision | |
| Feb 24, 2026 | 83.9 | MathVision | |
| Feb 24, 2026 | 86 | MathVision | |
| Feb 24, 2026 | 86.2 | MathVision | |
| Feb 16, 2026 | 88.6 | MathVision | |
| Feb 14, 2026 | 88.8 | MathVision | |
| Feb 6, 2026 | 74.3 | MathVision | |
| Jan 27, 2026 | Kimi K2.5 | 84.2 | MathVision |
| Jan 13, 2026 | 70.8 | MathVision (SeRe) | |
| Dec 18, 2025 | 81.3 | MathVision | |
| Nov 11, 2025 | 61.9 | MathVision | |
| Sep 23, 2025 | 74.6 | MathVision | |
| Sep 22, 2025 | 56.3 | MathVision_full | |
| Aug 25, 2025 | 63.9 | MathVision | |
| Aug 11, 2025 | GLM-4.5V | 65.6 | MathVision |
| Jul 31, 2025 | 64.8 | MATH-Vision | |
| Jul 24, 2025 | 62.5 | MathVision | |
| Jun 25, 2025 | 70.3 | MathVision | |
| Jun 21, 2025 | Kimi-VL-A3B-Thinking-2506 | 56.9 | MATH-Vision (Pass@1) |
| May 12, 2025 | 68.7 | MathVision | |
| Apr 10, 2025 | Kimi-VL-A3B-Thinking | 36.8 | MathVision (full) (Pass@1) |
| Mar 27, 2025 | 25 | MathVisionfull | |
| Jan 28, 2025 | 38.1 | MathVision_FULL | |
| Jan 20, 2025 | Kimi k1.5 | 38.6 | MathVision-Full (Pass@1), long-CoT |
32 system cards report it, from 12 labs
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| Nov 20, 2026 | | 89 | MathVision | Capabilities |
| Oct 23, 2026 | | 46.3 | MathVision, reasoning, 43.72 non-reasoning (Table 8) | Capabilities |
| Oct 15, 2026 | | 78.2 | MathVision (Table 10) | Capabilities |
| Sep 22, 2026 | | 91.8 | MathVision without CI, some ground truth corrected (Table 3). With CI 96.2 | Capabilities |
| Jul 27, 2026 | Kimi K3 technical report | 94.3 | Math-Vision without Python, cell 94.3 / 97.8 (Table 2) | Capabilities |
| Jul 24, 2026 | | 59.5 | MATH-Vision, max resolution, thinking (Table 6) | Capabilities |
| Jul 24, 2026 | | 52.4 | MATH-Vision, max resolution, thinking (Table 6) | Capabilities |
| Jul 24, 2026 | | 85.6 | MATH-Vision, thinking, max resolution (Table 6) | Capabilities |
| Jul 24, 2026 | | 82.4 | MATH-Vision, thinking, max resolution (Table 6) | Capabilities |
| Jun 23, 2026 | | 92.6 | MathVision, no tools, 94.5 with tools (Table 9) | Capabilities |
| Apr 17, 2026 | | 73 | MathVision (Table 6) | Capabilities |
| Apr 9, 2026 | | 75.2 | MATH-Vision (Table 2, Reasoning mode) | Capabilities |
| Mar 21, 2026 | | 81.3 | MathVision (Table 2) | Capabilities |
| Mar 4, 2026 | | 36.2 | MathVision MINI, default mode, 34.2 force nothink, 36.2 force thinking (Tables 5 and 6) | Capabilities |
| Feb 14, 2026 | | 88.8 | MathVision (Table 8) | Capabilities |
| Feb 4, 2026 | | 74.3 | MathVision (Table 3) | Capabilities |
| Jan 30, 2026 | Kimi K2.5 technical report | 84.2 | MathVision (Table 4) | Capabilities |
| Jan 14, 2026 | | 70.8 | MathVision, SeRe (Tables 1 and 3). PaCoRe 75.95 | Capabilities |
| Jan 1, 2026 | GLM-4.6V technical report | 63.5 | MathVision, GLM-4.6V thinking (Table 2) | Capabilities |
| Nov 26, 2025 | | 74.6 | MathVision (Table 2) | Capabilities |
| Nov 6, 2025 | | 53.6 | MathVision, reasoning on (Table 3). Reasoning off 31.5 | Capabilities |
| Sep 22, 2025 | | 56.3 | MATH-Vision full, Instruct (Table 9). Thinking 62.9 (Table 10) | Capabilities |
| Aug 25, 2025 | | 63.9 | MathVision (Table 2, Table 3). With parallel thinking 65.9 | Capabilities |
| Aug 21, 2025 | | 62.5 | MathVision (Table 2) | Capabilities |
| Aug 13, 2025 | GLM-4.5V technical report | 65.6 | MathVision, GLM-4.5V thinking (Table 2) | Capabilities |
| Jul 1, 2025 | GLM-4.1V-9B-Thinking technical report | 54.4 | MathVision, GLM-4.1V-9B-Thinking thinking (Table 2) | Capabilities |
| Jun 23, 2025 | Kimi-VL-A3B-Thinking-2506 technical report | 56.9 | MathVision (full) (Pass@1) (Table 4) | Capabilities |
| May 11, 2025 | | 68.7 | MathVision, thinking (Table 6). Non-thinking 65.5 | Capabilities |
| Apr 10, 2025 | Kimi-VL-A3B-Thinking technical report | 36.8 | MathVision (full) (Pass@1) (Table 4). Figure 13 shows 18.7 at 1k rising to 36.8 at 16k max thinking tokens | Capabilities |
| Mar 26, 2025 | | 25 | MathVision full (Table 5) | Capabilities |
| Feb 19, 2025 | | 38.1 | MATH-Vision full (Table 3) | Capabilities |
| Jan 20, 2025 | Kimi k1.5 technical report | 38.6 | MathVision-Full (Pass@1), long-CoT (Table 2). Short-CoT 31.0 (Table 3) | Capabilities |
The lab's number comes out ahead in 0 of 3 pairs that differ
Each score a lab reported for its own model, beside the same model's best score on the leaderboard. Gaps mostly come from different harnesses, effort levels and task subsets. 22 of them match, which usually means one side copied the other. Every benchmark.
| Model | Lab says | Board says | Gap |
|---|---|---|---|
| | 70.3% MathVision | 77.2% Leaderboard | −6.9 pts |
| Kimi K2.6 | 87.4% MathVision | 93.2% Leaderboard, with python | −5.8 pts |
| Kimi K2.5 | 84.2% MathVision | 85% Leaderboard, thinking, with python | −0.8 pts |
| Kimi k1.5 | 38.6% MathVision-Full (Pass@1), long-CoT | 38.6% Leaderboard | Same |
| | 38.1% MathVision_FULL | 38.1% Leaderboard | Same |
| Kimi-VL-A3B-Thinking | 36.8% MathVision (full) (Pass@1) | 36.8% Leaderboard | Same |
| | 68.7% MathVision | 68.7% Leaderboard | Same |
| Kimi-VL-A3B-Thinking-2506 | 56.9% MATH-Vision (Pass@1) | 56.9% Leaderboard | Same |
| | 64.8% MATH-Vision | 64.8% Leaderboard | Same |
| GLM-4.5V | 65.6% MathVision | 65.6% Leaderboard | Same |
| | 63.9% MathVision | 63.9% Leaderboard | Same |
| | 74.6% MathVision | 74.6% Leaderboard, thinking | Same |
| | 81.3% MathVision | 81.3% Leaderboard | Same |
| | 88.8% MathVision | 88.8% Leaderboard | Same |
| | 88.6% MathVision | 88.6% Leaderboard | Same |
| | 86.2% MathVision | 86.2% Leaderboard | Same |
| | 86% MathVision | 86% Leaderboard | Same |
| | 83.9% MathVision | 83.9% Leaderboard | Same |
| | 85.6% MATH-Vision | 85.6% Leaderboard | Same |
| | 82.4% MATH-Vision | 82.4% Leaderboard | Same |
| | 52.4% MATH-Vision | 52.4% Leaderboard | Same |
| | 59.5% MATH-Vision | 59.5% Leaderboard | Same |
| | 89.8% MathVision | 89.8% Leaderboard | Same |
| GLM-4.1V-9B-Thinking | 54.4% MathVision, GLM-4.1V-9B-Thinking thinking (Table 2) | 54.4% Leaderboard | Same |
| GLM-4.6V | 63.5% MathVision, GLM-4.6V thinking (Table 2) | 63.5% Leaderboard | Same |
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%
- 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%
- ChartQA In use Models answer human-written and generated questions about chart images that need visual and arithmetic reasoning. Scored by relaxed accuracy with a 5 percent numeric tolerance.
Command A Vision 90.9%
Data audit
All 10 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 (153 entries), launch posts (48 scores) and system cards (32 scores).
- Pass Models scored 133 models we track, plus 48 other entries on its leaderboard.
- Pass Newest score The newest is from November 2026, for Inkling 2 Mini.
- Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Kimi K3 (#25).
- Note Headroom Saturated. The best is 97.7%, past 85% since November 2025.
Leaderboard import 3 pass
- Pass Importer We read the board on mathllm.github.io: 111 rows for models we track and 48 other entries at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 2 pass
- Pass Lab claims All 25 pairs agree within 10 points.
- Pass Known issues No published audit found problems.