# Mathematical Proof Reasoning on IMO-ProofBench Advanced

Source: [Mathematical Proof Reasoning on IMO-ProofBench Advanced](<https://www.sota2.com/research/sota/mathematical-proof-reasoning-on-imo-proofbench-advanced>)

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

Data updated: 2026-06-16T07:59:29.903Z

Evaluation of model performance on mathematical proof reasoning using the Advanced subset of the IMO-ProofBench dataset, measured by Mask-Proof Avg@4 and Expert score.

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 Reasoning |
| Dataset (including any stated split) | IMO-ProofBench Advanced (full set) |
| Metric | Mask-Proof Average@4 |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 14 of 14 |
| 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

- [Mask-Proof Average@4](<https://www.sota2.com/research/sota/mathematical-proof-reasoning-on-imo-proofbench-advanced.md?metric=Mask-Proof%20Average%404>): Higher is better; unit: PERCENT.

- [Expert Score](<https://www.sota2.com/research/sota/mathematical-proof-reasoning-on-imo-proofbench-advanced.md?metric=Expert%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Mask-Proof Average@4 | Result date | Report paper |
| --- | --- | --- | --- |
| Gemini 2.5 Pro | 58.6 | Not recorded | 6a30d8bfd33dfb23e783cdad |
| o3 | 53.1 | Not recorded | 6a30d8bfd33dfb23e783cdad |
| Grok 4.1 Reasoning | 53.1 | Not recorded | 6a30d8bfd33dfb23e783cdad |
| o3 | 52.9 | Not recorded | 6a30d8bfd33dfb23e783cdad |
| Grok 4.1 Reasoning | 50 | Not recorded | 6a30d8bfd33dfb23e783cdad |
| Gemini 2.5 Pro | 46.9 | Not recorded | 6a30d8bfd33dfb23e783cdad |
| o4-mini | 46.9 | Not recorded | 6a30d8bfd33dfb23e783cdad |
| o4-mini | 42.1 | Not recorded | 6a30d8bfd33dfb23e783cdad |
| Qwen3-235B | 34.3 | Not recorded | 6a30d8bfd33dfb23e783cdad |
| DeepSeek V3 | 32.9 | Not recorded | 6a30d8bfd33dfb23e783cdad |
| Qwen3-235B | 31.2 | Not recorded | 6a30d8bfd33dfb23e783cdad |
| Kimi-K2 | 30.7 | Not recorded | 6a30d8bfd33dfb23e783cdad |
| Kimi-K2 | 25 | Not recorded | 6a30d8bfd33dfb23e783cdad |
| DeepSeek V3 | 25 | Not recorded | 6a30d8bfd33dfb23e783cdad |

### Gemini 2.5 Pro

Mask-Proof Average@4: 58.6. Result ID: 6a310218fbc373592c2d0d8d.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8bfd33dfb23e783cdad.md>)

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

### o3

Mask-Proof Average@4: 53.1. Result ID: 6a310218fbc373592c2d0d9a.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8bfd33dfb23e783cdad.md>)

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

### Grok 4.1 Reasoning

Mask-Proof Average@4: 53.1. Result ID: 6a310218fbc373592c2d0d9c.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8bfd33dfb23e783cdad.md>)

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

### o3

Mask-Proof Average@4: 52.9. Result ID: 6a310218fbc373592c2d0d8c.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8bfd33dfb23e783cdad.md>)

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

### Grok 4.1 Reasoning

Mask-Proof Average@4: 50. Result ID: 6a310218fbc373592c2d0d8e.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8bfd33dfb23e783cdad.md>)

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

### Gemini 2.5 Pro

Mask-Proof Average@4: 46.9. Result ID: 6a310218fbc373592c2d0d9b.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8bfd33dfb23e783cdad.md>)

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

### o4-mini

Mask-Proof Average@4: 46.9. Result ID: 6a310218fbc373592c2d0d9d.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8bfd33dfb23e783cdad.md>)

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

### o4-mini

Mask-Proof Average@4: 42.1. Result ID: 6a310218fbc373592c2d0d8f.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8bfd33dfb23e783cdad.md>)

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

### Qwen3-235B

Mask-Proof Average@4: 34.3. Result ID: 6a310218fbc373592c2d0d91.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8bfd33dfb23e783cdad.md>)

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

### DeepSeek V3

Mask-Proof Average@4: 32.9. Result ID: 6a310218fbc373592c2d0d92.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8bfd33dfb23e783cdad.md>)

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

### Qwen3-235B

Mask-Proof Average@4: 31.2. Result ID: 6a310218fbc373592c2d0d9f.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8bfd33dfb23e783cdad.md>)

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

### Kimi-K2

Mask-Proof Average@4: 30.7. Result ID: 6a310218fbc373592c2d0d90.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8bfd33dfb23e783cdad.md>)

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

### Kimi-K2

Mask-Proof Average@4: 25. Result ID: 6a310218fbc373592c2d0d9e.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8bfd33dfb23e783cdad.md>)

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

### DeepSeek V3

Mask-Proof Average@4: 25. Result ID: 6a310218fbc373592c2d0da0.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8bfd33dfb23e783cdad.md>)

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

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