# Reward Modeling on Overall 5-Benchmark Suite

Source: [Reward Modeling on Overall 5-Benchmark Suite](<https://www.sota2.com/research/sota/reward-modeling-on-overall-5-benchmark-suite>)

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

Data updated: 2026-04-10T07:19:22.337Z

Evaluation of model performance on the Overall 5-Benchmark Suite for the Reward Modeling task, reporting AVG and ∆ metrics.

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 | Reward Modeling |
| Dataset (including any stated split) | Overall 5-Benchmark Suite |
| Metric | Average Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 12 of 12 |
| 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

- [Average Score](<https://www.sota2.com/research/sota/reward-modeling-on-overall-5-benchmark-suite.md?metric=Average%20Score>): Higher is better; unit: PERCENT.

- [Score Delta](<https://www.sota2.com/research/sota/reward-modeling-on-overall-5-benchmark-suite.md?metric=Score%20Delta>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Average Score | Result date | Report paper |
| --- | --- | --- | --- |
| ConsistRM (Qwen3-8B) | 73.5 | Not recorded | 69d883df7d0172140ec777ee |
| Qwen3-8B + RFT | 72.5 | Not recorded | 69d883df7d0172140ec777ee |
| Qwen3-8B + CAAR | 72.2 | Not recorded | 69d883df7d0172140ec777ee |
| Qwen3-8B + TTRL | 71.7 | Not recorded | 69d883df7d0172140ec777ee |
| ConsistRM (Qwen3-4B) | 71 | Not recorded | 69d883df7d0172140ec777ee |
| Qwen3-8B | 70.9 | Not recorded | 69d883df7d0172140ec777ee |
| Qwen3-8B + SFT | 70.7 | Not recorded | 69d883df7d0172140ec777ee |
| Qwen3-4B + CAAR | 70.3 | Not recorded | 69d883df7d0172140ec777ee |
| Qwen3-4B + RFT | 69.8 | Not recorded | 69d883df7d0172140ec777ee |
| Qwen3-4B + TTRL | 69.5 | Not recorded | 69d883df7d0172140ec777ee |
| Qwen3-4B | 68.6 | Not recorded | 69d883df7d0172140ec777ee |
| Qwen3-4B + SFT | 67.8 | Not recorded | 69d883df7d0172140ec777ee |

### ConsistRM (Qwen3-8B)

Average Score: 73.5. Result ID: 69d8a46047f34ce924c118b4.

[Reporting paper](<https://www.sota2.com/research/paper/69d883df7d0172140ec777ee.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69d883df7d0172140ec777ee.md>)

Recorded evaluation configuration: {"GT":"False"}. This record does not assert independent reproduction.

### Qwen3-8B + RFT

Average Score: 72.5. Result ID: 69d8a46047f34ce924c118b1.

[Reporting paper](<https://www.sota2.com/research/paper/69d883df7d0172140ec777ee.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73101e209131cea5a395.md>)

[Implementation](<https://github.com/deepseek-ai/DeepSeek-Math>)

Recorded evaluation configuration: {"GT":"True"}. This record does not assert independent reproduction.

### Qwen3-8B + CAAR

Average Score: 72.2. Result ID: 69d8a46047f34ce924c118b3.

[Reporting paper](<https://www.sota2.com/research/paper/69d883df7d0172140ec777ee.md>)

Recorded evaluation configuration: {"GT":"False"}. This record does not assert independent reproduction.

### Qwen3-8B + TTRL

Average Score: 71.7. Result ID: 69d8a46047f34ce924c118b2.

[Reporting paper](<https://www.sota2.com/research/paper/69d883df7d0172140ec777ee.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79ce1e209131cea7f7aa.md>)

Recorded evaluation configuration: {"GT":"False"}. This record does not assert independent reproduction.

### ConsistRM (Qwen3-4B)

Average Score: 71. Result ID: 69d8a46047f34ce924c118ae.

[Reporting paper](<https://www.sota2.com/research/paper/69d883df7d0172140ec777ee.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69d883df7d0172140ec777ee.md>)

Recorded evaluation configuration: {"GT":"False"}. This record does not assert independent reproduction.

### Qwen3-8B

Average Score: 70.9. Result ID: 69d8a46047f34ce924c118af.

[Reporting paper](<https://www.sota2.com/research/paper/69d883df7d0172140ec777ee.md>)

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

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

### Qwen3-8B + SFT

Average Score: 70.7. Result ID: 69d8a46047f34ce924c118b0.

[Reporting paper](<https://www.sota2.com/research/paper/69d883df7d0172140ec777ee.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78a01e209131cea796d7.md>)

Recorded evaluation configuration: {"GT":"True"}. This record does not assert independent reproduction.

### Qwen3-4B + CAAR

Average Score: 70.3. Result ID: 69d8a46047f34ce924c118ad.

[Reporting paper](<https://www.sota2.com/research/paper/69d883df7d0172140ec777ee.md>)

Recorded evaluation configuration: {"GT":"False"}. This record does not assert independent reproduction.

### Qwen3-4B + RFT

Average Score: 69.8. Result ID: 69d8a46047f34ce924c118ab.

[Reporting paper](<https://www.sota2.com/research/paper/69d883df7d0172140ec777ee.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73101e209131cea5a395.md>)

[Implementation](<https://github.com/deepseek-ai/DeepSeek-Math>)

Recorded evaluation configuration: {"GT":"True"}. This record does not assert independent reproduction.

### Qwen3-4B + TTRL

Average Score: 69.5. Result ID: 69d8a46047f34ce924c118ac.

[Reporting paper](<https://www.sota2.com/research/paper/69d883df7d0172140ec777ee.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79ce1e209131cea7f7aa.md>)

Recorded evaluation configuration: {"GT":"False"}. This record does not assert independent reproduction.

### Qwen3-4B

Average Score: 68.6. Result ID: 69d8a46047f34ce924c118a9.

[Reporting paper](<https://www.sota2.com/research/paper/69d883df7d0172140ec777ee.md>)

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

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

### Qwen3-4B + SFT

Average Score: 67.8. Result ID: 69d8a46047f34ce924c118aa.

[Reporting paper](<https://www.sota2.com/research/paper/69d883df7d0172140ec777ee.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78a01e209131cea796d7.md>)

Recorded evaluation configuration: {"GT":"True"}. This record does not assert independent reproduction.

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