# Classification on MMLU (test)

Source: [Classification on MMLU (test)](<https://www.sota2.com/research/sota/classification-on-mmlu-test>)

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

Data updated: 2026-02-26T05:34:02.404Z

Evaluation of classification performance on the test split of the MMLU benchmark, reporting 0-shot and 5-shot accuracy results.

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 | Classification |
| Dataset (including any stated split) | MMLU (test) |
| Metric | Accuracy (5-shot) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 31 of 31 |
| 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 (5-shot)](<https://www.sota2.com/research/sota/classification-on-mmlu-test.md?metric=Accuracy%20(5-shot)>): Higher is better; unit: PERCENT.

- [Accuracy (0-shot)](<https://www.sota2.com/research/sota/classification-on-mmlu-test.md?metric=Accuracy%20(0-shot)>): Higher is better; unit: PERCENT.

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

## Recorded results

| Method | Accuracy (5-shot) | Result date | Report paper |
| --- | --- | --- | --- |
| OpenAI code-davinci-002 | 77.4 | Not recorded | 68fa6e791e209131cea401cb |
| OpenAI text-davinci-003 | 74.2 | Not recorded | 68fa6e791e209131cea401cb |
| FLAN-PaLM 540B | 73.5 | Not recorded | 68fa6e791e209131cea401cb |
| PaLM 540B | 71.3 | Not recorded | 68fa6e791e209131cea401cb |
| OpenAI text-davinci-002 | 64.5 | Not recorded | 68fa6e791e209131cea401cb |
| FLAN-PaLM 62B | 59.6 | Not recorded | 68fa6e791e209131cea401cb |
| PaLM 62B | 55.1 | Not recorded | 68fa6e791e209131cea401cb |
| FLAN-T5 11B | 54.9 | Not recorded | 68fa6e791e209131cea401cb |
| OPT-IML-Max 175B | 47.1 | Not recorded | 68fa6e791e209131cea401cb |
| OPT-IML-Max 30B | 43.2 | Not recorded | 68fa6e791e209131cea401cb |
| OPT 175B | 34.2 | Not recorded | 68fa6e791e209131cea401cb |
| T0pp 11B | 33.7 | Not recorded | 68fa6e791e209131cea401cb |
| OpenAI davinci | 32.3 | Not recorded | 68fa6e791e209131cea401cb |
| OPT-IML-Max 1.3B | 29.5 | Not recorded | 68fa6e791e209131cea401cb |
| OPT 30B | 26.1 | Not recorded | 68fa6e791e209131cea401cb |
| OPT 1.3B | 25.9 | Not recorded | 68fa6e791e209131cea401cb |
| T5 11B | 25.9 | Not recorded | 68fa6e791e209131cea401cb |

### OpenAI code-davinci-002

Accuracy (5-shot): 77.4. Result ID: 699023f73a6c314849fe64ff.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

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

### OpenAI text-davinci-003

Accuracy (5-shot): 74.2. Result ID: 699023f73a6c314849fe64fe.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

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

### FLAN-PaLM 540B

Accuracy (5-shot): 73.5. Result ID: 699023f73a6c314849fe64fc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e201e209131cea3dfa7.md>)

[Implementation](<https://github.com/google-research/t5x/blob/main/docs/models.md#flan-t5-checkpoints>)

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

### PaLM 540B

Accuracy (5-shot): 71.3. Result ID: 699023f73a6c314849fe64f4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d931e209131cea3abf2.md>)

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

### OpenAI text-davinci-002

Accuracy (5-shot): 64.5. Result ID: 699023f73a6c314849fe64fd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

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

### FLAN-PaLM 62B

Accuracy (5-shot): 59.6. Result ID: 699023f73a6c314849fe64fb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e201e209131cea3dfa7.md>)

[Implementation](<https://github.com/google-research/t5x/blob/main/docs/models.md#flan-t5-checkpoints>)

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

### PaLM 62B

Accuracy (5-shot): 55.1. Result ID: 699023f73a6c314849fe64f3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d931e209131cea3abf2.md>)

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

### FLAN-T5 11B

Accuracy (5-shot): 54.9. Result ID: 699023f73a6c314849fe64fa.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e201e209131cea3dfa7.md>)

[Implementation](<https://github.com/google-research/t5x/blob/main/docs/models.md#flan-t5-checkpoints>)

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

### OPT-IML-Max 175B

Accuracy (5-shot): 47.1. Result ID: 699023f73a6c314849fe64f8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

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

### OPT-IML-Max 30B

Accuracy (5-shot): 43.2. Result ID: 699023f73a6c314849fe64f7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

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

### OPT 175B

Accuracy (5-shot): 34.2. Result ID: 699023f73a6c314849fe64f1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cda1e209131cea36474.md>)

[Implementation](<https://github.com/facebookresearch/metaseq>)

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

### T0pp 11B

Accuracy (5-shot): 33.7. Result ID: 699023f73a6c314849fe64f9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c131e209131cea31f71.md>)

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

### OpenAI davinci

Accuracy (5-shot): 32.3. Result ID: 699023f73a6c314849fe64f5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa680c1e209131cea19d53.md>)

[Implementation](<https://github.com/openai/gpt-3/blob/master/overlap_frequency.md>)

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

### OPT-IML-Max 1.3B

Accuracy (5-shot): 29.5. Result ID: 699023f73a6c314849fe64f6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

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

### OPT 30B

Accuracy (5-shot): 26.1. Result ID: 699023f73a6c314849fe64f0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cda1e209131cea36474.md>)

[Implementation](<https://github.com/facebookresearch/metaseq>)

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

### OPT 1.3B

Accuracy (5-shot): 25.9. Result ID: 699023f73a6c314849fe64ef.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cda1e209131cea36474.md>)

[Implementation](<https://github.com/facebookresearch/metaseq>)

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

### T5 11B

Accuracy (5-shot): 25.9. Result ID: 699023f73a6c314849fe64f2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e791e209131cea401cb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa709e1e209131cea4c470.md>)

[Implementation](<https://github.com/google-research/text-to-text-transfer-transformer>)

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

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