# Mathematical Reasoning on MinervaMath (Acc, Tokens, EFF)

Source: [Mathematical Reasoning on MinervaMath (Acc, Tokens, EFF)](<https://www.sota2.com/research/sota/mathematical-reasoning-on-minervamath-acc-tokens-eff>)

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

Data updated: 2026-04-20T06:41:57.199Z

Evaluation of model performance on mathematical reasoning tasks using the MinervaMath dataset, reporting Accuracy (Acc), Token count, and Efficiency (EFF).

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) | MinervaMath |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 36 of 36 |
| 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-minervamath-acc-tokens-eff.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [Token Count](<https://www.sota2.com/research/sota/mathematical-reasoning-on-minervamath-acc-tokens-eff.md?metric=Token%20Count>): Higher is better; unit: SCALAR.

- [Efficiency](<https://www.sota2.com/research/sota/mathematical-reasoning-on-minervamath-acc-tokens-eff.md?metric=Efficiency>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| ThinkPrune(6k-\>3k) | 54.04 | Not recorded | 69a66bcb2d603e6de3e3cb9e |
| D-SFT-\>D-RL3k-\>D-RL6k | 53.68 | Not recorded | 69a66bcb2d603e6de3e3cb9e |
| D-SFT-\>D-RL6k | 53.68 | Not recorded | 69a66bcb2d603e6de3e3cb9e |
| ThinkPrune(6k) | 53.68 | Not recorded | 69a66bcb2d603e6de3e3cb9e |
| Original | 52.21 | Not recorded | 69a66bcb2d603e6de3e3cb9e |
| D-SFT-\>D-RL6k | 50 | Not recorded | 69a66bcb2d603e6de3e3cb9e |
| D-SFT-\>D-RL6k | 49.26 | Not recorded | 69a66bcb2d603e6de3e3cb9e |
| D-SFT-\>D-RL3k-\>D-RL6k | 48.53 | Not recorded | 69a66bcb2d603e6de3e3cb9e |
| REO-RL(Q-Spec)\* | 46.8 | Not recorded | 69a66bcb2d603e6de3e3cb9e |
| D-SFT-\>D-RL3k-\>D-RL6k | 46.32 | Not recorded | 69a66bcb2d603e6de3e3cb9e |
| D-SFT-\>D-RL3k | 44.49 | Not recorded | 69a66bcb2d603e6de3e3cb9e |
| D-SFT | 37.87 | Not recorded | 69a66bcb2d603e6de3e3cb9e |
| Mean Acc. | 34 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Self-Cert. | 34 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Loc. non-uni | 34 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Loc. uni | 34 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Glob. uni | 34 | Not recorded | 68fa7bed1e209131cea8b5ad |
| High Conf. | 33 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Low Ent. | 33 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Glob. non-uni | 33 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Loc. uni | 31 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Mean Acc. | 30 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Self-Cert. | 30 | Not recorded | 68fa7bed1e209131cea8b5ad |
| High Conf. | 30 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Low Ent. | 30 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Glob. non-uni | 30 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Glob. uni | 28 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Loc. non-uni | 26 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Mean Acc. | 23 | Not recorded | 68fa7bed1e209131cea8b5ad |
| High Conf. | 23 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Loc. non-uni | 23 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Loc. uni | 23 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Glob. uni | 23 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Self-Cert. | 22 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Low Ent. | 22 | Not recorded | 68fa7bed1e209131cea8b5ad |
| Glob. non-uni | 21 | Not recorded | 68fa7bed1e209131cea8b5ad |

### ThinkPrune(6k-\>3k)

Accuracy: 54.04. Result ID: 69a89bbd6dbb7075b9517f3e.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77c61e209131cea753e8.md>)

Recorded evaluation configuration: {"Prompt":"step","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### D-SFT-\>D-RL3k-\>D-RL6k

Accuracy: 53.68. Result ID: 69a89bbd6dbb7075b9517f38.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

Recorded evaluation configuration: {"Prompt":"step","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### D-SFT-\>D-RL6k

Accuracy: 53.68. Result ID: 69a89bbd6dbb7075b9517f3b.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

Recorded evaluation configuration: {"Prompt":"step","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### ThinkPrune(6k)

Accuracy: 53.68. Result ID: 69a89bbd6dbb7075b9517f3d.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77c61e209131cea753e8.md>)

Recorded evaluation configuration: {"Prompt":"step","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### Original

Accuracy: 52.21. Result ID: 69a89bbd6dbb7075b9517f34.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

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

Recorded evaluation configuration: {"Prompt":"step","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### D-SFT-\>D-RL6k

Accuracy: 50. Result ID: 69a89bbd6dbb7075b9517f3c.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

Recorded evaluation configuration: {"Prompt":"adaptive","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### D-SFT-\>D-RL6k

Accuracy: 49.26. Result ID: 69a89bbd6dbb7075b9517f3a.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

Recorded evaluation configuration: {"Prompt":"draft","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### D-SFT-\>D-RL3k-\>D-RL6k

Accuracy: 48.53. Result ID: 69a89bbd6dbb7075b9517f39.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

Recorded evaluation configuration: {"Prompt":"adaptive","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### REO-RL(Q-Spec)\*

Accuracy: 46.8. Result ID: 69a89bbd6dbb7075b9517f3f.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b411e209131cea8752e.md>)

Recorded evaluation configuration: {"Prompt":"step","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### D-SFT-\>D-RL3k-\>D-RL6k

Accuracy: 46.32. Result ID: 69a89bbd6dbb7075b9517f37.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

Recorded evaluation configuration: {"Prompt":"draft","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### D-SFT-\>D-RL3k

Accuracy: 44.49. Result ID: 69a89bbd6dbb7075b9517f36.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

Recorded evaluation configuration: {"Prompt":"draft","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### D-SFT

Accuracy: 37.87. Result ID: 69a89bbd6dbb7075b9517f35.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66bcb2d603e6de3e3cb9e.md>)

Recorded evaluation configuration: {"Prompt":"draft","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### Mean Acc.

Accuracy: 34. Result ID: 69e5ca9747f34ce924cb5914.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-8B"}. This record does not assert independent reproduction.

### Self-Cert.

Accuracy: 34. Result ID: 69e5ca9747f34ce924cb5915.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77291e209131cea717ab.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-8B"}. This record does not assert independent reproduction.

### Loc. non-uni

Accuracy: 34. Result ID: 69e5ca9747f34ce924cb5918.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-8B"}. This record does not assert independent reproduction.

### Loc. uni

Accuracy: 34. Result ID: 69e5ca9747f34ce924cb5919.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-8B"}. This record does not assert independent reproduction.

### Glob. uni

Accuracy: 34. Result ID: 69e5ca9747f34ce924cb591b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-8B"}. This record does not assert independent reproduction.

### High Conf.

Accuracy: 33. Result ID: 69e5ca9747f34ce924cb5916.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78b61e209131cea79f87.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-8B"}. This record does not assert independent reproduction.

### Low Ent.

Accuracy: 33. Result ID: 69e5ca9747f34ce924cb5917.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-8B"}. This record does not assert independent reproduction.

### Glob. non-uni

Accuracy: 33. Result ID: 69e5ca9747f34ce924cb591a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-8B"}. This record does not assert independent reproduction.

### Loc. uni

Accuracy: 31. Result ID: 69e5ca9747f34ce924cb5909.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction.

### Mean Acc.

Accuracy: 30. Result ID: 69e5ca9747f34ce924cb5904.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction.

### Self-Cert.

Accuracy: 30. Result ID: 69e5ca9747f34ce924cb5905.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77291e209131cea717ab.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction.

### High Conf.

Accuracy: 30. Result ID: 69e5ca9747f34ce924cb5906.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78b61e209131cea79f87.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction.

### Low Ent.

Accuracy: 30. Result ID: 69e5ca9747f34ce924cb5907.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction.

### Glob. non-uni

Accuracy: 30. Result ID: 69e5ca9747f34ce924cb590a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction.

### Glob. uni

Accuracy: 28. Result ID: 69e5ca9747f34ce924cb590b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction.

### Loc. non-uni

Accuracy: 26. Result ID: 69e5ca9747f34ce924cb5908.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction.

### Mean Acc.

Accuracy: 23. Result ID: 69e5ca9747f34ce924cb590c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Llama-8B"}. This record does not assert independent reproduction.

### High Conf.

Accuracy: 23. Result ID: 69e5ca9747f34ce924cb590e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78b61e209131cea79f87.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Llama-8B"}. This record does not assert independent reproduction.

### Loc. non-uni

Accuracy: 23. Result ID: 69e5ca9747f34ce924cb5910.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Llama-8B"}. This record does not assert independent reproduction.

### Loc. uni

Accuracy: 23. Result ID: 69e5ca9747f34ce924cb5911.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Llama-8B"}. This record does not assert independent reproduction.

### Glob. uni

Accuracy: 23. Result ID: 69e5ca9747f34ce924cb5913.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Llama-8B"}. This record does not assert independent reproduction.

### Self-Cert.

Accuracy: 22. Result ID: 69e5ca9747f34ce924cb590d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77291e209131cea717ab.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Llama-8B"}. This record does not assert independent reproduction.

### Low Ent.

Accuracy: 22. Result ID: 69e5ca9747f34ce924cb590f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Llama-8B"}. This record does not assert independent reproduction.

### Glob. non-uni

Accuracy: 21. Result ID: 69e5ca9747f34ce924cb5912.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bed1e209131cea8b5ad.md>)

Recorded evaluation configuration: {"Backbone Model":"DS-R1-Distill-Llama-8B"}. This record does not assert independent reproduction.

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