# 5-shot Performance on CEval, MMLU, GSM8K, Arc-E, Arc-C (test)

Source: [5-shot Performance on CEval, MMLU, GSM8K, Arc-E, Arc-C (test)](<https://www.sota2.com/research/sota/5-shot-performance-on-ceval-mmlu-gsm8k-arc-e-arc-c-test>)

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

Data updated: 2026-02-26T05:38:18.323Z

Evaluation reporting 5-shot performance across the CEval, MMLU, GSM8K, Arc-E, and Arc-C benchmarks on their respective test splits.

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 | Few-shot (5-shot) performance |
| Dataset (including any stated split) | CEval, MMLU, GSM8K, Arc-E, Arc-C (test) |
| Metric | CEval |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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

- [CEval](<https://www.sota2.com/research/sota/5-shot-performance-on-ceval-mmlu-gsm8k-arc-e-arc-c-test.md?metric=CEval>): Higher is better; unit: PERCENT.

- [MMLU](<https://www.sota2.com/research/sota/5-shot-performance-on-ceval-mmlu-gsm8k-arc-e-arc-c-test.md?metric=MMLU>): Higher is better; unit: PERCENT.

- [GSM8K](<https://www.sota2.com/research/sota/5-shot-performance-on-ceval-mmlu-gsm8k-arc-e-arc-c-test.md?metric=GSM8K>): Higher is better; unit: PERCENT.

- [Arc-E](<https://www.sota2.com/research/sota/5-shot-performance-on-ceval-mmlu-gsm8k-arc-e-arc-c-test.md?metric=Arc-E>): Higher is better; unit: PERCENT.

- [Arc-C](<https://www.sota2.com/research/sota/5-shot-performance-on-ceval-mmlu-gsm8k-arc-e-arc-c-test.md?metric=Arc-C>): Higher is better; unit: PERCENT.

- [Average Score](<https://www.sota2.com/research/sota/5-shot-performance-on-ceval-mmlu-gsm8k-arc-e-arc-c-test.md?metric=Average%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | CEval | Result date | Report paper |
| --- | --- | --- | --- |
| Qwen-14B | 79.86 | Not recorded | 68fa74471e209131cea60b5e |
| DDK (Ours) | 65.38 | Not recorded | 68fa74471e209131cea60b5e |
| Qwen-1.5 (Student) + TED | 62.11 | Not recorded | 68fa74471e209131cea60b5e |
| Qwen-1.5 (Student) + MiniLLM | 62.03 | Not recorded | 68fa74471e209131cea60b5e |
| Qwen-1.5 (Student) | 61.96 | Not recorded | 68fa74471e209131cea60b5e |
| Qwen-1.5 (Student) + KD | 61.66 | Not recorded | 68fa74471e209131cea60b5e |
| Qwen-1.5 (Student) + CPT | 60.92 | Not recorded | 68fa74471e209131cea60b5e |

### Qwen-14B

CEval: 79.86. Result ID: 69934d423a6c3148491508b2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74471e209131cea60b5e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70b91e209131cea4cdc5.md>)

Recorded evaluation configuration: {"Few-shot":"5-shot","Model Size":"14B"}. This record does not assert independent reproduction.

### DDK (Ours)

CEval: 65.38. Result ID: 69934d423a6c3148491508b8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74471e209131cea60b5e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74471e209131cea60b5e.md>)

[Implementation](<https://github.com/microsoft/DeepSpeedExamples/tree/master/applications/DeepSpeed-Chat>)

Recorded evaluation configuration: {"Few-shot":"5-shot","Model Size":"1.8B","Strategy":"DDK"}. This record does not assert independent reproduction.

### Qwen-1.5 (Student) + TED

CEval: 62.11. Result ID: 69934d423a6c3148491508b6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74471e209131cea60b5e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fa11e209131cea46c8e.md>)

Recorded evaluation configuration: {"Few-shot":"5-shot","Model Size":"1.8B","Strategy":"TED"}. This record does not assert independent reproduction.

### Qwen-1.5 (Student) + MiniLLM

CEval: 62.03. Result ID: 69934d423a6c3148491508b7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74471e209131cea60b5e.md>)

Recorded evaluation configuration: {"Few-shot":"5-shot","Model Size":"1.8B","Strategy":"MiniLLM"}. This record does not assert independent reproduction.

### Qwen-1.5 (Student)

CEval: 61.96. Result ID: 69934d423a6c3148491508b3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74471e209131cea60b5e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70b91e209131cea4cdc5.md>)

Recorded evaluation configuration: {"Few-shot":"5-shot","Model Size":"1.8B"}. This record does not assert independent reproduction.

### Qwen-1.5 (Student) + KD

CEval: 61.66. Result ID: 69934d423a6c3148491508b5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74471e209131cea60b5e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa61b81e209131ce9f5b87.md>)

Recorded evaluation configuration: {"Few-shot":"5-shot","Model Size":"1.8B","Strategy":"KD"}. This record does not assert independent reproduction.

### Qwen-1.5 (Student) + CPT

CEval: 60.92. Result ID: 69934d423a6c3148491508b4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74471e209131cea60b5e.md>)

Recorded evaluation configuration: {"Few-shot":"5-shot","Model Size":"1.8B","Strategy":"CPT"}. This record does not assert independent reproduction.

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