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
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
| # | Model | Score |
|---|---|---|
| 1 | Doubao Seed 2.1 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.7% |
| 2 | Doubao Seed 2.1 Turbo· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.5% |
| 3 | Doubao Seed 1.8· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 87.7% |
| 4 | OpenAI o3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 86.8% |
| 5 | o4 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 84.3% |
| 6 | GPT 5· OpenVLM Leaderboard data | 81.9% |
| 7 | Gemini 2.5 Pro· OpenVLM Leaderboard data | 80.9% |
| 8 | Command A Plus 05 2026· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 80.6% |
| 9 | GPT-5 mini· OpenVLM Leaderboard data | 79.2% |
| 10 | InternVL3 78B· OpenVLM Leaderboard data | 79% |
| 11 | SenseNova-V6-Pro· OpenVLM Leaderboard data | 76.9% |
| 12 | internvl3-38b· OpenVLM Leaderboard data | 76.3% |
| 13 | Qwen2.5 VL 72B Instruct· OpenVLM Leaderboard data | 74.2% |
| 14 | Llama 4 Maverick· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 73.7% |
| 15 | GPT-4.1 mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 73.1% |
| 16 | GPT-5 nano· OpenVLM Leaderboard data | 73.1% |
| 17 | GPT-4.5 Preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 72.3% |
| 18 | GPT 4.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 72.2% |
| 19 | o1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 71.8% |
| 20 | ChatGPT-4o Latest· OpenVLM Leaderboard data | 71.6% |
| 21 | QVQ 72B Preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 71.4% |
| 22 | Llama 4 Scout· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 70.7% |
| 23 | Llama 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% |
| 24 | internvl3-8b· OpenVLM Leaderboard data | 70.5% |
| 25 | Gemini 2.0 Flash· OpenVLM Leaderboard data | 70.4% |
| 26 | Qwen2-VL-72B· OpenVLM Leaderboard data | 69.7% |
| 27 | Pixtral Large· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 69.4% |
| 28 | Grok 2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 69% |
| 29 | GLM 4 Vision Plus· OpenVLM Leaderboard data | 68.8% |
| 30 | Gemini 1.5 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 68.1% |
| 31 | Grok 2 mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 68.1% |
| 32 | Qwen2.5-VL 7B Instruct· OpenVLM Leaderboard data | 68.1% |
| 33 | gemini-1.5-pro-002· OpenVLM Leaderboard data | 67.9% |
| 34 | Qwen2.5-Omni 7B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 67.9% |
| 35 | Claude 3.5 Sonnet· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 67.7% |
| 36 | Mistral Small 3.2 24B Instruct2506· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 67.1% |
| 37 | Moonshot V1 8K· OpenVLM Leaderboard data | 66.9% |
| 38 | Grok 2 Vision 1212· OpenVLM Leaderboard data | 66.6% |
| 39 | Hunyuan Standard Vision· OpenVLM Leaderboard data | 66.1% |
| 40 | Gemini 1.5 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 65.8% |
| 41 | Phi 4 Multimodal Instruct· OpenVLM Leaderboard dataValues from 2 sources
| 65.8% |
| 42 | DeepSeek VL2· OpenVLM Leaderboard dataValues from 2 sources
| 63.9% |
| 43 | gemini-1.5-flash-002· OpenVLM Leaderboard data | 63.7% |
| 44 | GPT 4o· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 61.4% |
| 45 | qwen2.5-vl-3b-instruct· OpenVLM Leaderboard data | 61.2% |
| 46 | MiniCPM-V 2.6· OpenVLM Leaderboard data | 60.8% |
| 47 | Gemma 3 27B IT· OpenVLM Leaderboard data | 59.8% |
| 48 | InternVL2 26B· OpenVLM Leaderboard data | 59.6% |
| 49 | InternVL2 4B· OpenVLM Leaderboard data | 58.5% |
| 50 | Llama 3.2 90B Vision Instruct· OpenVLM Leaderboard data | 58.2% |
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.