# Broad Knowledge Question Answering on MMLU-Pro Source: [Broad Knowledge Question Answering on MMLU-Pro]() Collection index: [llms.txt]() Data updated: 2026-04-14T08:30:54.547Z Evaluation of model performance on the Broad Knowledge Question Answering task using the MMLU-Pro dataset, measured using Pass@1. 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 | Broad Knowledge Question Answering | | Dataset (including any stated split) | MMLU-Pro | | Metric | Pass@1 | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 10 of 10 | | 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 - [Pass@1](): Higher is better; unit: PERCENT. ## Recorded results | Method | Pass@1 | Result date | Report paper | | --- | --- | --- | --- | | Qwen2.5-7B | 36.1 | Not recorded | 68fa7bc11e209131cea8a35b | | Qwen2.5-7B + PPO w/ VERL. | 34.4 | Not recorded | 68fa7bc11e209131cea8a35b | | Qwen2.5-7B + PPO | 34.1 | Not recorded | 68fa7bc11e209131cea8a35b | | Qwen2.5-7B + GRPO | 33.8 | Not recorded | 68fa7bc11e209131cea8a35b | | Qwen2.5-7B + GRPO w/ VERL. | 33.4 | Not recorded | 68fa7bc11e209131cea8a35b | | Mathstral-7B-v0.1 + GRPO w/ VERL. | 30.7 | Not recorded | 68fa7bc11e209131cea8a35b | | Mathstral-7B-v0.1 + GRPO | 29.8 | Not recorded | 68fa7bc11e209131cea8a35b | | Mathstral-7B-v0.1 + PPO w/ VERL. | 29.3 | Not recorded | 68fa7bc11e209131cea8a35b | | Mathstral-7B-v0.1 + PPO | 29.1 | Not recorded | 68fa7bc11e209131cea8a35b | | Mathstral-7B-v0.1 | 26.4 | Not recorded | 68fa7bc11e209131cea8a35b | ### Qwen2.5-7B Pass@1: 36.1. Result ID: 69ddfb2147f34ce924c562d2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen2.5-7B"}. This record does not assert independent reproduction. ### Qwen2.5-7B + PPO w/ VERL. Pass@1: 34.4. Result ID: 69ddfb2147f34ce924c562d6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen2.5-7B","Training":"PPO w/ VERL."}. This record does not assert independent reproduction. ### Qwen2.5-7B + PPO Pass@1: 34.1. Result ID: 69ddfb2147f34ce924c562d5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen2.5-7B","Training":"PPO"}. This record does not assert independent reproduction. ### Qwen2.5-7B + GRPO Pass@1: 33.8. Result ID: 69ddfb2147f34ce924c562d3. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Base Model":"Qwen2.5-7B","Training":"GRPO"}. This record does not assert independent reproduction. ### Qwen2.5-7B + GRPO w/ VERL. Pass@1: 33.4. Result ID: 69ddfb2147f34ce924c562d4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen2.5-7B","Training":"GRPO w/ VERL."}. This record does not assert independent reproduction. ### Mathstral-7B-v0.1 + GRPO w/ VERL. Pass@1: 30.7. Result ID: 69ddfb2147f34ce924c562cf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Mathstral-7B-v0.1","Training":"GRPO w/ VERL."}. This record does not assert independent reproduction. ### Mathstral-7B-v0.1 + GRPO Pass@1: 29.8. Result ID: 69ddfb2147f34ce924c562ce. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Base Model":"Mathstral-7B-v0.1","Training":"GRPO"}. This record does not assert independent reproduction. ### Mathstral-7B-v0.1 + PPO w/ VERL. Pass@1: 29.3. Result ID: 69ddfb2147f34ce924c562d1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Mathstral-7B-v0.1","Training":"PPO w/ VERL."}. This record does not assert independent reproduction. ### Mathstral-7B-v0.1 + PPO Pass@1: 29.1. Result ID: 69ddfb2147f34ce924c562d0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Mathstral-7B-v0.1","Training":"PPO"}. This record does not assert independent reproduction. ### Mathstral-7B-v0.1 Pass@1: 26.4. Result ID: 69ddfb2147f34ce924c562cd. [Reporting paper]() Recorded evaluation configuration: {"Base Model":"Mathstral-7B-v0.1"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.