math benchmark
MathVista Leaderboard
MathVista evaluates the mathematical reasoning of foundation models in visual contexts using 6,141 examples.
Updated Sep 5, 2026
Models39
Model coverage39
MetricScore
EvidenceB
MathVista Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelBY | Score90.70% | Percentile100.00% | Participants39 | EvidenceC | Evaluated |
| Rank02 | ModelBY | Score90.50% | Percentile97.37% | Participants39 | EvidenceC | Evaluated |
| Rank03 | ModelOP | Score86.80% | Percentile94.74% | Participants39 | EvidenceC | Evaluated |
| Rank04 | ModelOP | Score84.30% | Percentile92.11% | Participants39 | EvidenceC | Evaluated |
| Rank05 | ModelST | Score84.00% | Percentile89.47% | Participants39 | EvidenceC | Evaluated |
| Rank06 | ModelCO | Score80.60% | Percentile86.84% | Participants39 | EvidenceC | Evaluated |
| Rank07 | ModelMA | Score74.90% | Percentile84.21% | Participants39 | EvidenceC | Evaluated |
| Rank08 | ModelME | Score73.70% | Percentile81.58% | Participants39 | EvidenceC | Evaluated |
| Rank09 | ModelOP | Score73.10% | Percentile78.95% | Participants39 | EvidenceC | Evaluated |
| Rank10 | ModelOP | Score72.30% | Percentile76.32% | Participants39 | EvidenceC | Evaluated |
| Rank11 | ModelOP | Score72.20% | Percentile73.68% | Participants39 | EvidenceC | Evaluated |
| Rank12 | ModelOP | Score71.80% | Percentile71.05% | Participants39 | EvidenceC | Evaluated |
| Rank13 | ModelAC | Score71.40% | Percentile68.42% | Participants39 | EvidenceC | Evaluated |
| Rank14 | ModelME | Score70.70% | Percentile65.79% | Participants39 | EvidenceC | Evaluated |
| Rank15 | ModelMA | Score69.40% | Percentile63.16% | Participants39 | EvidenceC | Evaluated |
| Rank16 | ModelXA | Score69.00% | Percentile60.53% | Participants39 | EvidenceC | Evaluated |
| Rank17 | ModelGO | Score68.10% | Percentile57.89% | Participants39 | EvidenceC | Evaluated |
| Rank18 | ModelXA | Score68.10% | Percentile55.26% | Participants39 | EvidenceC | Evaluated |
| Rank19 | ModelAC | Score67.90% | Percentile52.63% | Participants39 | EvidenceC | Evaluated |
| Rank20 | ModelAN | Score67.70% | Percentile50.00% | Participants39 | EvidenceC | Evaluated |
| Rank21 | ModelMA | Score67.09% | Percentile47.37% | Participants39 | EvidenceC | Evaluated |
| Rank22 | ModelGO | Score65.80% | Percentile44.74% | Participants39 | EvidenceC | Evaluated |
| Rank23 | ModelOP | Score63.80% | Percentile42.11% | Participants39 | EvidenceC | Evaluated |
| Rank24 | ModelDE | Score62.80% | Percentile39.47% | Participants39 | EvidenceC | Evaluated |
| Rank25 | ModelMI | Score62.40% | Percentile36.84% | Participants39 | EvidenceC | Evaluated |
| Rank26 | ModelOP | Score61.40% | Percentile34.21% | Participants39 | EvidenceC | Evaluated |
| Rank27 | ModelDE | Score60.70% | Percentile31.58% | Participants39 | EvidenceC | Evaluated |
| Rank28 | ModelMA | Score58.00% | Percentile28.95% | Participants39 | EvidenceC | Evaluated |
| Rank29 | ModelME | Score57.30% | Percentile26.32% | Participants39 | EvidenceC | Evaluated |
| Rank30 | ModelOP | Score56.70% | Percentile23.68% | Participants39 | EvidenceC | Evaluated |
MathVista Highlights
The leading models and scores on this benchmark.
MathVista Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for MathVista
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is MathVista?
What MathVista measures and how its scores work.
MathVista is a benchmark comprising 6,141 examples derived from 28 existing multimodal datasets and three newly created datasets: IQTest, FunctionQA, and PaperQA.
It measures models' ability to understand complex figures and perform rigorous reasoning across diverse mathematical and visual tasks, reported as a Score ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- MathVista
- Modality
- multimodal
- Primary category
- math
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about MathVista.
Which model scores highest on MathVista?
Seed 2.1 Pro is currently ranked first with 90.70%.
What are the top three models on MathVista?
The current leaders are Seed 2.1 Pro (90.70%), Seed 2.1 Turbo (90.50%), and o3 (86.80%).
Which MathVista model has the lowest official input price?
Qwen2.5-Omni-7B has the lowest matched official input price at $0.10 input / $0.40 output per 1M tokens.
Which models are fastest among MathVista results?
The fastest matched records are Llama 4 Scout (776.10 tok/s via Groq), GPT-4.1 mini (467.39 tok/s via OpenAI), and Llama 4 Maverick (307.30 tok/s via Groq).
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.