# Mathematical Reasoning on APEX 2025

Source: [Mathematical Reasoning on APEX 2025](<https://www.sota2.com/research/sota/mathematical-reasoning-on-apex-2025>)

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

Data updated: 2026-05-08T11:01:27.077Z

Evaluation of model performance on the APEX 2025 dataset for mathematical reasoning tasks, typically measured using accuracy.

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 Reasoning |
| Dataset (including any stated split) | APEX 2025 |
| Metric | Accuracy |
| 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

- [Accuracy](<https://www.sota2.com/research/sota/mathematical-reasoning-on-apex-2025.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| TRICE-30B | 16.7 | Not recorded | 69fd6e555497b3b670c89694 |
| TRICE-4B | 13.9 | Not recorded | 69fd6e555497b3b670c89694 |
| GLM-4.7-Flash w/ recipe | 13.9 | Not recorded | 69fd6e555497b3b670c89694 |
| GLM-4.7-Flash | 11.1 | Not recorded | 69fd6e555497b3b670c89694 |
| Nemotron-3-Nano-30B-A3B | 11.1 | Not recorded | 69fd6e555497b3b670c89694 |
| GPT-OSS-20B | 8.3 | Not recorded | 69fd6e555497b3b670c89694 |
| TRICE-4B | 5.6 | Not recorded | 69fd6e555497b3b670c89694 |
| Qwen3-4B-Thinking-2507 | 2.8 | Not recorded | 69fd6e555497b3b670c89694 |
| Qwen3-30B-A3B-Instruct-2507 | 2.8 | Not recorded | 69fd6e555497b3b670c89694 |
| Qwen3.5-4B | 0 | Not recorded | 69fd6e555497b3b670c89694 |
| Qwen3.5-9B | 0 | Not recorded | 69fd6e555497b3b670c89694 |
| Qwen3-30B-A3B-Thinking-2507 | 0 | Not recorded | 69fd6e555497b3b670c89694 |
| Qwen3.5-35B-A3B | 0 | Not recorded | 69fd6e555497b3b670c89694 |
| TRICE-30B | 0 | Not recorded | 69fd6e555497b3b670c89694 |

### TRICE-30B

Accuracy: 16.7. Result ID: 69fdc27b47f34ce924e18a2e.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

Recorded evaluation configuration: {"Tool Usage":"true","Parameter Scale":"~30B","Evaluation Protocol":"unified"}. This record does not assert independent reproduction.

### TRICE-4B

Accuracy: 13.9. Result ID: 69fdc27b47f34ce924e18a25.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

Recorded evaluation configuration: {"Tool Usage":"true","Parameter Scale":"\<10B","Evaluation Protocol":"unified"}. This record does not assert independent reproduction.

### GLM-4.7-Flash w/ recipe

Accuracy: 13.9. Result ID: 69fdc27b47f34ce924e18a2d.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a891e209131cea83c41.md>)

Recorded evaluation configuration: {"Tool Usage":"true","Parameter Scale":"~30B","Evaluation Protocol":"unified"}. This record does not assert independent reproduction.

### GLM-4.7-Flash

Accuracy: 11.1. Result ID: 69fdc27b47f34ce924e18a2b.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a891e209131cea83c41.md>)

Recorded evaluation configuration: {"Tool Usage":"true","Parameter Scale":"~30B","Evaluation Protocol":"unified"}. This record does not assert independent reproduction.

### Nemotron-3-Nano-30B-A3B

Accuracy: 11.1. Result ID: 69fdc27b47f34ce924e18a2c.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

Recorded evaluation configuration: {"Tool Usage":"true","Parameter Scale":"~30B","Evaluation Protocol":"unified"}. This record does not assert independent reproduction.

### GPT-OSS-20B

Accuracy: 8.3. Result ID: 69fdc27b47f34ce924e18a2a.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ab01e209131cea84a0c.md>)

Recorded evaluation configuration: {"Tool Usage":"true","Parameter Scale":"~30B","Evaluation Protocol":"unified"}. This record does not assert independent reproduction.

### TRICE-4B

Accuracy: 5.6. Result ID: 69fdc27b47f34ce924e18a24.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

Recorded evaluation configuration: {"Tool Usage":"false","Parameter Scale":"\<10B","Evaluation Protocol":"unified"}. This record does not assert independent reproduction.

### Qwen3-4B-Thinking-2507

Accuracy: 2.8. Result ID: 69fdc27b47f34ce924e18a21.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Tool Usage":"false","Parameter Scale":"\<10B","Evaluation Protocol":"unified"}. This record does not assert independent reproduction.

### Qwen3-30B-A3B-Instruct-2507

Accuracy: 2.8. Result ID: 69fdc27b47f34ce924e18a26.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Tool Usage":"false","Parameter Scale":"~30B","Evaluation Protocol":"unified"}. This record does not assert independent reproduction.

### Qwen3.5-4B

Accuracy: 0. Result ID: 69fdc27b47f34ce924e18a22.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

Recorded evaluation configuration: {"Tool Usage":"false","Parameter Scale":"\<10B","Evaluation Protocol":"unified"}. This record does not assert independent reproduction.

### Qwen3.5-9B

Accuracy: 0. Result ID: 69fdc27b47f34ce924e18a23.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

Recorded evaluation configuration: {"Tool Usage":"false","Parameter Scale":"\<10B","Evaluation Protocol":"unified"}. This record does not assert independent reproduction.

### Qwen3-30B-A3B-Thinking-2507

Accuracy: 0. Result ID: 69fdc27b47f34ce924e18a27.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Tool Usage":"false","Parameter Scale":"~30B","Evaluation Protocol":"unified"}. This record does not assert independent reproduction.

### Qwen3.5-35B-A3B

Accuracy: 0. Result ID: 69fdc27b47f34ce924e18a28.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

Recorded evaluation configuration: {"Tool Usage":"false","Parameter Scale":"~30B","Evaluation Protocol":"unified"}. This record does not assert independent reproduction.

### TRICE-30B

Accuracy: 0. Result ID: 69fdc27b47f34ce924e18a29.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69fd6e555497b3b670c89694.md>)

Recorded evaluation configuration: {"Tool Usage":"false","Parameter Scale":"~30B","Evaluation Protocol":"unified"}. This record does not assert independent reproduction.

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