math benchmark
MathVision Leaderboard
MATH-Vision is a dataset designed to measure multimodal mathematical reasoning capabilities. It focuses on evaluating how well models can solve mathematical problems that require both visual understanding and mathematical reasoning, bridging the gap between visual and mathematical domains.
Updated Aug 17, 2026
Models33
Model coverage33
MetricScore
EvidenceB
MathVision Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelMA | Score97.8% | Percentile100.0% | Participants33 | EvidenceC | Evaluated |
| Rank02 | ModelAC | Score94.6% | Percentile96.9% | Participants33 | EvidenceC | Evaluated |
| Rank03 | ModelBY | Score94.5% | Percentile93.8% | Participants33 | EvidenceC | Evaluated |
| Rank04 | ModelMA | Score93.2% | Percentile90.6% | Participants33 | EvidenceC | Evaluated |
| Rank05 | ModelBY | Score92.7% | Percentile87.5% | Participants33 | EvidenceC | Evaluated |
| Rank06 | ModelAC | Score90.3% | Percentile84.4% | Participants33 | EvidenceC | Evaluated |
| Rank07 | ModelAC | Score88.0% | Percentile81.3% | Participants33 | EvidenceC | Evaluated |
| Rank08 | ModelAC | Score86.2% | Percentile78.1% | Participants33 | EvidenceC | Evaluated |
| Rank09 | ModelAC | Score86.0% | Percentile75.0% | Participants33 | EvidenceC | Evaluated |
| Rank10 | ModelGO | Score85.6% | Percentile71.9% | Participants33 | EvidenceC | Evaluated |
| Rank11 | ModelMA | Score84.2% | Percentile68.8% | Participants33 | EvidenceC | Evaluated |
| Rank12 | ModelAC | Score83.9% | Percentile65.6% | Participants33 | EvidenceC | Evaluated |
| Rank13 | ModelGO | Score82.4% | Percentile62.5% | Participants33 | EvidenceC | Evaluated |
| Rank14 | ModelGO | Score79.7% | Percentile59.4% | Participants33 | EvidenceC | Evaluated |
| Rank15 | ModelAC | Score74.6% | Percentile56.3% | Participants33 | EvidenceC | Evaluated |
| Rank16 | ModelST | Score70.8% | Percentile53.1% | Participants33 | EvidenceC | Evaluated |
| Rank17 | ModelGO | Score70.5% | Percentile50.0% | Participants33 | EvidenceC | Evaluated |
| Rank18 | ModelAC | Score70.2% | Percentile46.9% | Participants33 | EvidenceC | Evaluated |
| Rank19 | ModelAC | Score66.5% | Percentile43.8% | Participants33 | EvidenceC | Evaluated |
| Rank20 | ModelAC | Score65.7% | Percentile40.6% | Participants33 | EvidenceC | Evaluated |
| Rank21 | ModelAC | Score63.4% | Percentile37.5% | Participants33 | EvidenceC | Evaluated |
| Rank22 | ModelAC | Score62.7% | Percentile34.4% | Participants33 | EvidenceC | Evaluated |
| Rank23 | ModelAC | Score60.2% | Percentile31.3% | Participants33 | EvidenceC | Evaluated |
| Rank24 | ModelAC | Score60.0% | Percentile28.1% | Participants33 | EvidenceC | Evaluated |
| Rank25 | ModelGO | Score59.5% | Percentile25.0% | Participants33 | EvidenceC | Evaluated |
| Rank26 | ModelAC | Score53.9% | Percentile21.9% | Participants33 | EvidenceC | Evaluated |
| Rank27 | ModelGO | Score52.4% | Percentile18.8% | Participants33 | EvidenceC | Evaluated |
| Rank28 | ModelAC | Score51.6% | Percentile15.6% | Participants33 | EvidenceC | Evaluated |
| Rank29 | ModelAC | Score38.4% | Percentile12.5% | Participants33 | EvidenceC | Evaluated |
| Rank30 | ModelAC | Score38.1% | Percentile9.4% | Participants33 | EvidenceC | Evaluated |
| Rank31 | ModelAC | Score35.9% | Percentile6.3% | Participants33 | EvidenceC | Evaluated |
| Rank32 | ModelAC | Score25.1% | Percentile3.1% | Participants33 | EvidenceC | Evaluated |
| Rank33 | ModelAC | Score25.0% | Percentile0.0% | Participants33 | EvidenceC | Evaluated |
MathVision Highlights
The leading models and scores on this benchmark.
MathVision Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for MathVision
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis mathvision AI model leaderboard uses descending score in the benchmark's original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
- 01MAKimi K3Moonshot AI
- Score
- 97.8%
- Price
- $3.0 input / $15 output per 1M tokens
- Speed
- Up to 26 tok/s via Fireworks
Strengths
- Ranks #1 of 33 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures MathVision, not total model capability
- 02ACQwen3.8-27BAlibaba Cloud / Qwen Team
- Score
- 94.6%
Strengths
- Ranks #2 of 33 compared models
- 97th percentile on this benchmark
- C evidence result
Considerations
- This result measures MathVision, not total model capability
- 03BYSeed 2.1 ProByteDance
- Score
- 94.5%
Strengths
- Ranks #3 of 33 compared models
- 94th percentile on this benchmark
- C evidence result
Considerations
- This result measures MathVision, not total model capability
- 04MAKimi K2.6Moonshot AI
- Score
- 93.2%
- Price
- $0.95 input / $4.0 output per 1M tokens
- Speed
- Up to 285 tok/s via Fireworks
Strengths
- Ranks #4 of 33 compared models
- 91th percentile on this benchmark
- C evidence result
Considerations
- This result measures MathVision, not total model capability
- 05BYSeed 2.1 TurboByteDance
- Score
- 92.7%
Strengths
- Ranks #5 of 33 compared models
- 88th percentile on this benchmark
- C evidence result
Considerations
- This result measures MathVision, not total model capability
Selection summary
Best AI Models for MathVision
Kimi K3 currently leads MathVision with 97.8%. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price and runtime.
Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.
What is MathVision?
What MathVision measures and how its scores work.
MATH-Vision is a dataset designed to measure multimodal mathematical reasoning capabilities. It focuses on evaluating how well models can solve mathematical problems that require both visual understanding and mathematical reasoning, bridging the gap between visual and mathematical domains.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- MathVision
- Modality
- multimodal
- Primary category
- math
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- mathvision|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about MathVision.
Which model scores highest on MathVision?
Kimi K3 is currently ranked first with 97.8%.
What does MathVision measure?
MATH-Vision is a dataset designed to measure multimodal mathematical reasoning capabilities. It focuses on evaluating how well models can solve mathematical problems that require both visual understanding and mathematical reasoning, bridging the gap between visual and mathematical domains.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
33 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.