# Medical Question Answering on HealthBench Hard 1000 Source: [Medical Question Answering on HealthBench Hard 1000]() Collection index: [llms.txt]() Data updated: 2026-05-08T10:20:33.938Z Evaluation of model performance on the HealthBench Hard 1000 subset for Medical Question Answering, 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 | Medical Question Answering | | Dataset (including any stated split) | HealthBench Hard 1000 | | Metric | Accuracy | | 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 - [Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Claude Sonnet 4.6 | 86 | Not recorded | 69fd6e1a5497b3b670c89462 | | Gemini 3.1 Pro | 80.8 | Not recorded | 69fd6e1a5497b3b670c89462 | | Gemini 3 Flash | 80.7 | Not recorded | 69fd6e1a5497b3b670c89462 | | Claude Opus 4.6 | 80.2 | Not recorded | 69fd6e1a5497b3b670c89462 | | Claude Opus 4.5 | 78.9 | Not recorded | 69fd6e1a5497b3b670c89462 | | Claude Opus 4.6 | 76.2 | Not recorded | 69fd6e1a5497b3b670c89462 | | Claude Sonnet 4.5 | 75.3 | Not recorded | 69fd6e1a5497b3b670c89462 | | Claude Opus 4.5 | 72.5 | Not recorded | 69fd6e1a5497b3b670c89462 | | Gemini 3.1 Pro | 72 | Not recorded | 69fd6e1a5497b3b670c89462 | | Claude Sonnet 4.6 | 69.3 | Not recorded | 69fd6e1a5497b3b670c89462 | | Claude Sonnet 4.5 | 67.9 | Not recorded | 69fd6e1a5497b3b670c89462 | | Gemini 3 Flash | 59.7 | Not recorded | 69fd6e1a5497b3b670c89462 | ### Claude Sonnet 4.6 Accuracy: 86. Result ID: 69fdb8ea47f34ce924e14752. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"Commercial LLMs","Setting":"BioMedArena"}. This record does not assert independent reproduction. ### Gemini 3.1 Pro Accuracy: 80.8. Result ID: 69fdb8ea47f34ce924e1474e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"Commercial LLMs","Setting":"BioMedArena"}. This record does not assert independent reproduction. ### Gemini 3 Flash Accuracy: 80.7. Result ID: 69fdb8ea47f34ce924e1474c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"Commercial LLMs","Setting":"BioMedArena"}. This record does not assert independent reproduction. ### Claude Opus 4.6 Accuracy: 80.2. Result ID: 69fdb8ea47f34ce924e14756. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"Commercial LLMs","Setting":"BioMedArena"}. This record does not assert independent reproduction. ### Claude Opus 4.5 Accuracy: 78.9. Result ID: 69fdb8ea47f34ce924e14754. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"Commercial LLMs","Setting":"BioMedArena"}. This record does not assert independent reproduction. ### Claude Opus 4.6 Accuracy: 76.2. Result ID: 69fdb8ea47f34ce924e14755. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"Commercial LLMs","Setting":"Baseline"}. This record does not assert independent reproduction. ### Claude Sonnet 4.5 Accuracy: 75.3. Result ID: 69fdb8ea47f34ce924e14750. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"Commercial LLMs","Setting":"BioMedArena"}. This record does not assert independent reproduction. ### Claude Opus 4.5 Accuracy: 72.5. Result ID: 69fdb8ea47f34ce924e14753. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"Commercial LLMs","Setting":"Baseline"}. This record does not assert independent reproduction. ### Gemini 3.1 Pro Accuracy: 72. Result ID: 69fdb8ea47f34ce924e1474d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"Commercial LLMs","Setting":"Baseline"}. This record does not assert independent reproduction. ### Claude Sonnet 4.6 Accuracy: 69.3. Result ID: 69fdb8ea47f34ce924e14751. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"Commercial LLMs","Setting":"Baseline"}. This record does not assert independent reproduction. ### Claude Sonnet 4.5 Accuracy: 67.9. Result ID: 69fdb8ea47f34ce924e1474f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"Commercial LLMs","Setting":"Baseline"}. This record does not assert independent reproduction. ### Gemini 3 Flash Accuracy: 59.7. Result ID: 69fdb8ea47f34ce924e1474b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"Commercial LLMs","Setting":"Baseline"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.