Skip to content

Benchmark

MathVista

Math benchmarks measure how well models solve maths problems. Scores are percentages. Higher is better.

Models measured
71
Scores recorded
73
Sources
3
IQ weight
1.22

Score distribution

62.7%
Mean
66.1%
Median
14.8%
Std dev
45%
10th percentile
79.2%
90th percentile
0%
Min
0%90.7%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.22 for chat models, measured over 70 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Doubao Seed 2.1 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim90.7%
2Doubao Seed 2.1 Turbo· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim90.5%
3Doubao Seed 1.8· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim87.7%
4OpenAI o3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim86.8%
5o4 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim84.3%
6GPT 5· OpenVLM Leaderboard data81.9%
7Gemini 2.5 Pro· OpenVLM Leaderboard data80.9%
8Command A Plus 05 2026· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim80.6%
9GPT-5 mini· OpenVLM Leaderboard data79.2%
10InternVL3 78B· OpenVLM Leaderboard data79%
11SenseNova-V6-Pro· OpenVLM Leaderboard data76.9%
12internvl3-38b· OpenVLM Leaderboard data76.3%
13Qwen2.5 VL 72B Instruct· OpenVLM Leaderboard data74.2%
14Llama 4 Maverick· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim73.7%
15GPT-4.1 mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
73.1%
16GPT-5 nano· OpenVLM Leaderboard data73.1%
17GPT-4.5 Preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
72.3%
18GPT 4.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
72.2%
19o1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim71.8%
20ChatGPT-4o Latest· OpenVLM Leaderboard data71.6%
21QVQ 72B Preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim71.4%
22Llama 4 Scout· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim70.7%
23Llama 4 Scoutinstruction-tuned table· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.70.7%
24internvl3-8b· OpenVLM Leaderboard data70.5%
25Gemini 2.0 Flash· OpenVLM Leaderboard data70.4%
26Qwen2-VL-72B· OpenVLM Leaderboard data69.7%
27Pixtral Large· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim69.4%
28Grok 2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim69%
29GLM 4 Vision Plus· OpenVLM Leaderboard data68.8%
30Gemini 1.5 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
68.1%
31Grok 2 mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim68.1%
32Qwen2.5-VL 7B Instruct· OpenVLM Leaderboard data68.1%
33gemini-1.5-pro-002· OpenVLM Leaderboard data67.9%
34Qwen2.5-Omni 7B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim67.9%
35Claude 3.5 Sonnet· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim67.7%
36Mistral Small 3.2 24B Instruct2506· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim67.1%
37Moonshot V1 8K· OpenVLM Leaderboard data66.9%
38Grok 2 Vision 1212· OpenVLM Leaderboard data66.6%
39Hunyuan Standard Vision· OpenVLM Leaderboard data66.1%
40Gemini 1.5 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
65.8%
41Phi 4 Multimodal Instruct· OpenVLM Leaderboard data
Values from 2 sources
65.8%
42DeepSeek VL2· OpenVLM Leaderboard data
Values from 2 sources
63.9%
43gemini-1.5-flash-002· OpenVLM Leaderboard data63.7%
44GPT 4o· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim61.4%
45qwen2.5-vl-3b-instruct· OpenVLM Leaderboard data61.2%
46MiniCPM-V 2.6· OpenVLM Leaderboard data60.8%
47Gemma 3 27B IT· OpenVLM Leaderboard data59.8%
48InternVL2 26B· OpenVLM Leaderboard data59.6%
49InternVL2 4B· OpenVLM Leaderboard data58.5%
50Llama 3.2 90B Vision Instruct· OpenVLM Leaderboard data58.2%

Showing 1–50 of 73 results.

Data credits: OpenVLM Leaderboard (OpenCompass / VLMEvalKit), Apache-2.0 (licence) · Data by LLM Stats · BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.