# Mathematical Reasoning on MATH 500 (Accuracy and Latency)

Source: [Mathematical Reasoning on MATH 500 (Accuracy and Latency)](<https://www.sota2.com/research/sota/mathematical-reasoning-on-math-500-accuracy-and-latency>)

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

Data updated: 2026-02-26T05:34:33.302Z

Evaluation of model performance on mathematical reasoning problems using the MATH 500 dataset, reporting pass@1 accuracy and relative latency.

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) | MATH 500 |
| Metric | Accuracy (pass@1) |
| 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 (pass@1)](<https://www.sota2.com/research/sota/mathematical-reasoning-on-math-500-accuracy-and-latency.md?metric=Accuracy%20(pass%401)>): Higher is better; unit: PERCENT.

- [Latency (relative)](<https://www.sota2.com/research/sota/mathematical-reasoning-on-math-500-accuracy-and-latency.md?metric=Latency%20(relative)>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Accuracy (pass@1) | Result date | Report paper |
| --- | --- | --- | --- |
| Power-SMC | 76.2 | Not recorded | 698d5e70c97a0b3e3b5e3ce5 |
| Scalable Power Sampling | 75.8 | Not recorded | 698d5e70c97a0b3e3b5e3ce5 |
| MH power sampling | 74.8 | Not recorded | 698d5e70c97a0b3e3b5e3ce5 |
| Power-SMC | 71.4 | Not recorded | 698d5e70c97a0b3e3b5e3ce5 |
| Scalable Power Sampling | 70.8 | Not recorded | 698d5e70c97a0b3e3b5e3ce5 |
| MH power sampling | 70.6 | Not recorded | 698d5e70c97a0b3e3b5e3ce5 |
| Low-temperature decoding | 69 | Not recorded | 698d5e70c97a0b3e3b5e3ce5 |
| Low-temperature decoding | 62.8 | Not recorded | 698d5e70c97a0b3e3b5e3ce5 |
| Power-SMC | 51.6 | Not recorded | 698d5e70c97a0b3e3b5e3ce5 |
| MH power sampling | 50.8 | Not recorded | 698d5e70c97a0b3e3b5e3ce5 |
| Baseline decoding | 49.8 | Not recorded | 698d5e70c97a0b3e3b5e3ce5 |
| Baseline decoding | 49.6 | Not recorded | 698d5e70c97a0b3e3b5e3ce5 |
| Low-temperature decoding | 47.8 | Not recorded | 698d5e70c97a0b3e3b5e3ce5 |
| Baseline decoding | 40 | Not recorded | 698d5e70c97a0b3e3b5e3ce5 |

### Power-SMC

Accuracy (pass@1): 76.2. Result ID: 699a2a4e5698665ec4afc8ef.

[Reporting paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

Recorded evaluation configuration: {"Model":"Qwen2.5-Math-7B"}. This record does not assert independent reproduction.

### Scalable Power Sampling

Accuracy (pass@1): 75.8. Result ID: 699a2a4e5698665ec4afc8ee.

[Reporting paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697c3b53a8b01e76cc8da5fa.md>)

Recorded evaluation configuration: {"Model":"Qwen2.5-Math-7B"}. This record does not assert independent reproduction.

### MH power sampling

Accuracy (pass@1): 74.8. Result ID: 699a2a4e5698665ec4afc8ed.

[Reporting paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c2a1e209131cea8cca6.md>)

Recorded evaluation configuration: {"Model":"Qwen2.5-Math-7B"}. This record does not assert independent reproduction.

### Power-SMC

Accuracy (pass@1): 71.4. Result ID: 699a2a4e5698665ec4afc8ea.

[Reporting paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

Recorded evaluation configuration: {"Model":"Qwen2.5-7B"}. This record does not assert independent reproduction.

### Scalable Power Sampling

Accuracy (pass@1): 70.8. Result ID: 699a2a4e5698665ec4afc8e9.

[Reporting paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697c3b53a8b01e76cc8da5fa.md>)

Recorded evaluation configuration: {"Model":"Qwen2.5-7B"}. This record does not assert independent reproduction.

### MH power sampling

Accuracy (pass@1): 70.6. Result ID: 699a2a4e5698665ec4afc8e8.

[Reporting paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c2a1e209131cea8cca6.md>)

Recorded evaluation configuration: {"Model":"Qwen2.5-7B"}. This record does not assert independent reproduction.

### Low-temperature decoding

Accuracy (pass@1): 69. Result ID: 699a2a4e5698665ec4afc8ec.

[Reporting paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

Recorded evaluation configuration: {"Model":"Qwen2.5-Math-7B","temperature":"τ=1/α"}. This record does not assert independent reproduction.

### Low-temperature decoding

Accuracy (pass@1): 62.8. Result ID: 699a2a4e5698665ec4afc8e7.

[Reporting paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

Recorded evaluation configuration: {"Model":"Qwen2.5-7B","temperature":"τ=1/α"}. This record does not assert independent reproduction.

### Power-SMC

Accuracy (pass@1): 51.6. Result ID: 699a2a4e5698665ec4afc8f3.

[Reporting paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

Recorded evaluation configuration: {"Model":"Phi-3.5-mini-instruct"}. This record does not assert independent reproduction.

### MH power sampling

Accuracy (pass@1): 50.8. Result ID: 699a2a4e5698665ec4afc8f2.

[Reporting paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c2a1e209131cea8cca6.md>)

Recorded evaluation configuration: {"Model":"Phi-3.5-mini-instruct"}. This record does not assert independent reproduction.

### Baseline decoding

Accuracy (pass@1): 49.8. Result ID: 699a2a4e5698665ec4afc8e6.

[Reporting paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

Recorded evaluation configuration: {"Model":"Qwen2.5-7B"}. This record does not assert independent reproduction.

### Baseline decoding

Accuracy (pass@1): 49.6. Result ID: 699a2a4e5698665ec4afc8eb.

[Reporting paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

Recorded evaluation configuration: {"Model":"Qwen2.5-Math-7B"}. This record does not assert independent reproduction.

### Low-temperature decoding

Accuracy (pass@1): 47.8. Result ID: 699a2a4e5698665ec4afc8f1.

[Reporting paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

Recorded evaluation configuration: {"Model":"Phi-3.5-mini-instruct","temperature":"τ=1/α"}. This record does not assert independent reproduction.

### Baseline decoding

Accuracy (pass@1): 40. Result ID: 699a2a4e5698665ec4afc8f0.

[Reporting paper](<https://www.sota2.com/research/paper/698d5e70c97a0b3e3b5e3ce5.md>)

Recorded evaluation configuration: {"Model":"Phi-3.5-mini-instruct"}. This record does not assert independent reproduction.

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