# General Knowledge Exam on MMLU-Pro (Accuracy)

Source: [General Knowledge Exam on MMLU-Pro (Accuracy)](<https://www.sota2.com/research/sota/general-knowledge-exam-on-mmlu-pro-accuracy>)

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

Data updated: 2026-02-26T05:50:02.583Z

Evaluation of model performance on the MMLU-Pro benchmark using the General Knowledge Exam task, measured by Accuracy.

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 | General Knowledge Exam |
| Dataset (including any stated split) | MMLU-Pro |
| Metric | MMLU-Pro Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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

- [MMLU-Pro Accuracy](<https://www.sota2.com/research/sota/general-knowledge-exam-on-mmlu-pro-accuracy.md?metric=MMLU-Pro%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | MMLU-Pro Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Gemini-2.5 (Pro) | 0.8645 | Not recorded | 69687481e25ad2fa9d1ec510 |
| Qwen3-VL (Thinking) | 0.8375 | Not recorded | 69687481e25ad2fa9d1ec510 |
| Seed-1.5-VL (Thinking) | 0.8339 | Not recorded | 69687481e25ad2fa9d1ec510 |
| GLM-4.6V | 0.7996 | Not recorded | 69687481e25ad2fa9d1ec510 |
| STEP3-VL-10B (PaCoRe) | 0.7709 | Not recorded | 69687481e25ad2fa9d1ec510 |
| STEP3-VL-10B (SeRe) | 0.7602 | Not recorded | 69687481e25ad2fa9d1ec510 |

### Gemini-2.5 (Pro)

MMLU-Pro Accuracy: 0.8645. Result ID: 6997bd365698665ec413d454.

[Reporting paper](<https://www.sota2.com/research/paper/69687481e25ad2fa9d1ec510.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c311e209131cea8cef5.md>)

Recorded evaluation configuration: {"Model Tier":"Pro"}. This record does not assert independent reproduction.

### Qwen3-VL (Thinking)

MMLU-Pro Accuracy: 0.8375. Result ID: 6997bd365698665ec413d453.

[Reporting paper](<https://www.sota2.com/research/paper/69687481e25ad2fa9d1ec510.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"Thinking Mode":"true","Parameters":"235B-A22B"}. This record does not assert independent reproduction.

### Seed-1.5-VL (Thinking)

MMLU-Pro Accuracy: 0.8339. Result ID: 6997bd365698665ec413d455.

[Reporting paper](<https://www.sota2.com/research/paper/69687481e25ad2fa9d1ec510.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78791e209131cea788a1.md>)

Recorded evaluation configuration: {"Thinking Mode":"true"}. This record does not assert independent reproduction.

### GLM-4.6V

MMLU-Pro Accuracy: 0.7996. Result ID: 6997bd365698665ec413d452.

[Reporting paper](<https://www.sota2.com/research/paper/69687481e25ad2fa9d1ec510.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7aaf1e209131cea849ba.md>)

Recorded evaluation configuration: {"Parameters":"106B-A12B"}. This record does not assert independent reproduction.

### STEP3-VL-10B (PaCoRe)

MMLU-Pro Accuracy: 0.7709. Result ID: 6997bd365698665ec413d451.

[Reporting paper](<https://www.sota2.com/research/paper/69687481e25ad2fa9d1ec510.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69687481e25ad2fa9d1ec510.md>)

[Implementation](<https://github.com/stepfun-ai/Step3-VL-10B>)

Recorded evaluation configuration: {"Reasoning Strategy":"PaCoRe","Parameters":"10B"}. This record does not assert independent reproduction.

### STEP3-VL-10B (SeRe)

MMLU-Pro Accuracy: 0.7602. Result ID: 6997bd365698665ec413d450.

[Reporting paper](<https://www.sota2.com/research/paper/69687481e25ad2fa9d1ec510.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69687481e25ad2fa9d1ec510.md>)

[Implementation](<https://github.com/stepfun-ai/Step3-VL-10B>)

Recorded evaluation configuration: {"Reasoning Strategy":"SeRe","Parameters":"10B"}. This record does not assert independent reproduction.

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