math benchmark
PolyMATH Leaderboard
Polymath is a challenging multi-modal mathematical reasoning benchmark designed to evaluate the general cognitive reasoning abilities of Multi-modal Large Language Models (MLLMs). The benchmark comprises 5,000 manually collected high-quality images of cognitive textual and visual challenges across 10 distinct categories, including pattern recognition, spatial reasoning, and relative reasoning.
Updated Aug 17, 2026
Models23
Model coverage23
MetricScore
EvidenceB
PolyMATH Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelAC | Score86.5% | Percentile100.0% | Participants23 | EvidenceC | Evaluated |
| Rank02 | ModelAC | Score84.0% | Percentile95.5% | Participants23 | EvidenceC | Evaluated |
| Rank03 | ModelAC | Score77.4% | Percentile90.9% | Participants23 | EvidenceC | Evaluated |
| Rank04 | ModelAC | Score73.3% | Percentile86.4% | Participants23 | EvidenceC | Evaluated |
| Rank05 | ModelAC | Score71.2% | Percentile81.8% | Participants23 | EvidenceC | Evaluated |
| Rank06 | ModelAC | Score68.9% | Percentile77.3% | Participants23 | EvidenceC | Evaluated |
| Rank07 | ModelAC | Score64.4% | Percentile72.7% | Participants23 | EvidenceC | Evaluated |
| Rank08 | ModelAC | Score60.1% | Percentile68.2% | Participants23 | EvidenceC | Evaluated |
| Rank09 | ModelAC | Score57.3% | Percentile63.6% | Participants23 | EvidenceC | Evaluated |
| Rank10 | ModelAC | Score56.3% | Percentile59.1% | Participants23 | EvidenceC | Evaluated |
| Rank11 | ModelAC | Score52.0% | Percentile54.5% | Participants23 | EvidenceC | Evaluated |
| Rank12 | ModelAC | Score51.7% | Percentile50.0% | Participants23 | EvidenceC | Evaluated |
| Rank13 | ModelAC | Score51.1% | Percentile45.5% | Participants23 | EvidenceC | Evaluated |
| Rank14 | ModelAC | Score50.2% | Percentile40.9% | Participants23 | EvidenceC | Evaluated |
| Rank15 | ModelAC | Score47.5% | Percentile36.4% | Participants23 | EvidenceC | Evaluated |
| Rank16 | ModelAC | Score45.9% | Percentile31.8% | Participants23 | EvidenceC | Evaluated |
| Rank17 | ModelAC | Score44.6% | Percentile27.3% | Participants23 | EvidenceC | Evaluated |
| Rank18 | ModelAC | Score44.3% | Percentile22.7% | Participants23 | EvidenceC | Evaluated |
| Rank19 | ModelAC | Score40.5% | Percentile18.2% | Participants23 | EvidenceC | Evaluated |
| Rank20 | ModelAC | Score30.4% | Percentile13.6% | Participants23 | EvidenceC | Evaluated |
| Rank21 | ModelAC | Score28.8% | Percentile9.1% | Participants23 | EvidenceC | Evaluated |
| Rank22 | ModelAC | Score26.1% | Percentile4.5% | Participants23 | EvidenceC | Evaluated |
| Rank23 | ModelAC | Score8.2% | Percentile0.0% | Participants23 | EvidenceC | Evaluated |
PolyMATH Highlights
The leading models and scores on this benchmark.
PolyMATH Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for PolyMATH
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis polymath 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.
- 01ACQwen3.7 MaxAlibaba Cloud / Qwen Team
- Score
- 86.5%
- Price
- $2.5 input / $7.5 output per 1M tokens
- Speed
- Up to 5.8 tok/s via Together
Strengths
- Ranks #1 of 23 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures PolyMATH, not total model capability
- 02ACQwen3.7-PlusAlibaba Cloud / Qwen Team
- Score
- 84.0%
- Price
- $0.50 input / $3.0 output per 1M tokens
Strengths
- Ranks #2 of 23 compared models
- 95th percentile on this benchmark
- C evidence result
Considerations
- This result measures PolyMATH, not total model capability
- 03ACQwen3.6 PlusAlibaba Cloud / Qwen Team
- Score
- 77.4%
- Price
- $0.50 input / $3.0 output per 1M tokens
- Speed
- Up to 16 tok/s via Together
Strengths
- Ranks #3 of 23 compared models
- 91th percentile on this benchmark
- C evidence result
Considerations
- This result measures PolyMATH, not total model capability
- 04ACQwen3.5-397B-A17BAlibaba Cloud / Qwen Team
- Score
- 73.3%
- Price
- $0.60 input / $3.6 output per 1M tokens
Strengths
- Ranks #4 of 23 compared models
- 86th percentile on this benchmark
- C evidence result
Considerations
- This result measures PolyMATH, not total model capability
- 05ACQwen3.5-27BAlibaba Cloud / Qwen Team
- Score
- 71.2%
- Price
- $0.30 input / $2.4 output per 1M tokens
- Speed
- Up to 6.7 tok/s via Novita
Strengths
- Ranks #5 of 23 compared models
- 82th percentile on this benchmark
- C evidence result
Considerations
- This result measures PolyMATH, not total model capability
Selection summary
Best AI Models for PolyMATH
Qwen3.7 Max currently leads PolyMATH with 86.5%. 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 PolyMATH?
What PolyMATH measures and how its scores work.
Polymath is a challenging multi-modal mathematical reasoning benchmark designed to evaluate the general cognitive reasoning abilities of Multi-modal Large Language Models (MLLMs). The benchmark comprises 5,000 manually collected high-quality images of cognitive textual and visual challenges across 10 distinct categories, including pattern recognition, spatial reasoning, and relative reasoning.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- PolyMATH
- Modality
- multimodal
- Primary category
- math
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- polymath|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about PolyMATH.
Which model scores highest on PolyMATH?
Qwen3.7 Max is currently ranked first with 86.5%.
What does PolyMATH measure?
Polymath is a challenging multi-modal mathematical reasoning benchmark designed to evaluate the general cognitive reasoning abilities of Multi-modal Large Language Models (MLLMs). The benchmark comprises 5,000 manually collected high-quality images of cognitive textual and visual challenges across 10 distinct categories, including pattern recognition, spatial reasoning, and relative reasoning.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
23 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.