# Mathematical Proof on USAMO 2026

Source: [Mathematical Proof on USAMO 2026](<https://www.sota2.com/research/sota/mathematical-proof-on-usamo-2026>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-05-20T08:03:50.353Z

Evaluation of model performance on mathematical proof tasks using the USAMO 2026 dataset, reporting overall percentage scores.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Mathematical Proof |
| Dataset (including any stated split) | USAMO 2026 |
| Metric | Overall Score (%) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 9 of 9 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Overall Score (%)](<https://www.sota2.com/research/sota/mathematical-proof-on-usamo-2026.md?metric=Overall%20Score%20(%25)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Overall Score (%) | Result date | Report paper |
| --- | --- | --- | --- |
| STAR-PólyaMath | 99.4 | Not recorded | 6a0d4002f984739ed6737770 |
| GPT-5.5 | 98.21 | Not recorded | 6a0d4002f984739ed6737770 |
| GPT-5.4 | 95.24 | Not recorded | 6a0d4002f984739ed6737770 |
| Gemini 3.1 Pro | 74.4 | Not recorded | 6a0d4002f984739ed6737770 |
| Claude Opus 4.7 | 63.1 | Not recorded | 6a0d4002f984739ed6737770 |
| DeepSeek-v4-Pro | 60.71 | Not recorded | 6a0d4002f984739ed6737770 |
| Kimi K2.6 | 51.19 | Not recorded | 6a0d4002f984739ed6737770 |
| GPT-5.2 | 50 | Not recorded | 6a0d4002f984739ed6737770 |
| Claude Opus 4.6 | 47.02 | Not recorded | 6a0d4002f984739ed6737770 |

### STAR-PólyaMath

Overall Score (%): 99.4. Result ID: 6a0d6a09fbc373592c0546e8.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4002f984739ed6737770.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0d4002f984739ed6737770.md>)

Recorded evaluation configuration: {"Backbone":"all GPT-5.5xh"}. This record does not assert independent reproduction.

### GPT-5.5

Overall Score (%): 98.21. Result ID: 6a0d6a09fbc373592c0546e0.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4002f984739ed6737770.md>)

Recorded evaluation configuration: {"Reasoning Effort":"xhigh"}. This record does not assert independent reproduction.

### GPT-5.4

Overall Score (%): 95.24. Result ID: 6a0d6a09fbc373592c0546e1.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4002f984739ed6737770.md>)

Recorded evaluation configuration: {"Reasoning Effort":"xhigh"}. This record does not assert independent reproduction.

### Gemini 3.1 Pro

Overall Score (%): 74.4. Result ID: 6a0d6a09fbc373592c0546e3.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4002f984739ed6737770.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Claude Opus 4.7

Overall Score (%): 63.1. Result ID: 6a0d6a09fbc373592c0546e4.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4002f984739ed6737770.md>)

Recorded evaluation configuration: {"Reasoning Effort":"xhigh"}. This record does not assert independent reproduction.

### DeepSeek-v4-Pro

Overall Score (%): 60.71. Result ID: 6a0d6a09fbc373592c0546e6.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4002f984739ed6737770.md>)

Recorded evaluation configuration: {"Reasoning Effort":"Max"}. This record does not assert independent reproduction.

### Kimi K2.6

Overall Score (%): 51.19. Result ID: 6a0d6a09fbc373592c0546e7.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4002f984739ed6737770.md>)

Recorded evaluation configuration: {"Reasoning Effort":"Think"}. This record does not assert independent reproduction.

### GPT-5.2

Overall Score (%): 50. Result ID: 6a0d6a09fbc373592c0546e2.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4002f984739ed6737770.md>)

Recorded evaluation configuration: {"Reasoning Effort":"high"}. This record does not assert independent reproduction.

### Claude Opus 4.6

Overall Score (%): 47.02. Result ID: 6a0d6a09fbc373592c0546e5.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4002f984739ed6737770.md>)

Recorded evaluation configuration: {"Reasoning Effort":"high"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.