# Clinical Reasoning on HealthBench Professional (525 cases) Source: [Clinical Reasoning on HealthBench Professional (525 cases)]() Collection index: [llms.txt]() Data updated: 2026-05-26T09:08:36.877Z Evaluation of model performance on the Clinical Reasoning task using the specific 525-case subset of the HealthBench Professional dataset. 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 | Clinical Reasoning | | Dataset (including any stated split) | HealthBench Professional (525 cases) | | Metric | Overall Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 11 of 11 | | 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 - [Overall Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Overall Score | Result date | Report paper | | --- | --- | --- | --- | | MDIA v1.0.53 | 62.72 | Not recorded | 6a152939dbd7150506f7e098 | | MDIA v1.0.50 | 61.66 | Not recorded | 6a152939dbd7150506f7e098 | | ChatGPT for Clinicians | 59 | Not recorded | 6a152939dbd7150506f7e098 | | MDIA v1.0.41 | 57.75 | Not recorded | 6a152939dbd7150506f7e098 | | GPT-5.4 base | 48.1 | Not recorded | 6a152939dbd7150506f7e098 | | Claude Opus 4.7 | 47 | Not recorded | 6a152939dbd7150506f7e098 | | GPT-5 | 46.2 | Not recorded | 6a152939dbd7150506f7e098 | | GPT-5.2 | 45.9 | Not recorded | 6a152939dbd7150506f7e098 | | Gemini 3.1 Pro | 43.8 | Not recorded | 6a152939dbd7150506f7e098 | | Physician-written baseline | 43.7 | Not recorded | 6a152939dbd7150506f7e098 | | Grok 4.20 | 36.1 | Not recorded | 6a152939dbd7150506f7e098 | ### MDIA v1.0.53 Overall Score: 62.72. Result ID: 6a15624dfbc373592c0cff4f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Description":"Hydra Platform, length-guided synthesizer + verifier","Avg len":"2789","Reference":"this work"}. This record does not assert independent reproduction. ### MDIA v1.0.50 Overall Score: 61.66. Result ID: 6a15624dfbc373592c0cff50. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Description":"Hydra Platform, Gemini 3.1 Pro, graph endpoint","Avg len":"4383","Reference":"this work"}. This record does not assert independent reproduction. ### ChatGPT for Clinicians Overall Score: 59. Result ID: 6a15624dfbc373592c0cff52. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Note":"best in OpenAI paper","Reference":"OpenAI 2026"}. This record does not assert independent reproduction. ### MDIA v1.0.41 Overall Score: 57.75. Result ID: 6a15624dfbc373592c0cff51. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Description":"Hydra Platform, single-agent endpoint","Reference":"this work"}. This record does not assert independent reproduction. ### GPT-5.4 base Overall Score: 48.1. Result ID: 6a15624dfbc373592c0cff53. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reference":"OpenAI 2026"}. This record does not assert independent reproduction. ### Claude Opus 4.7 Overall Score: 47. Result ID: 6a15624dfbc373592c0cff54. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reference":"OpenAI 2026"}. This record does not assert independent reproduction. ### GPT-5 Overall Score: 46.2. Result ID: 6a15624dfbc373592c0cff55. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reference":"OpenAI 2026"}. This record does not assert independent reproduction. ### GPT-5.2 Overall Score: 45.9. Result ID: 6a15624dfbc373592c0cff56. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reference":"OpenAI 2026"}. This record does not assert independent reproduction. ### Gemini 3.1 Pro Overall Score: 43.8. Result ID: 6a15624dfbc373592c0cff57. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reference":"OpenAI 2026"}. This record does not assert independent reproduction. ### Physician-written baseline Overall Score: 43.7. Result ID: 6a15624dfbc373592c0cff58. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reference":"OpenAI 2026"}. This record does not assert independent reproduction. ### Grok 4.20 Overall Score: 36.1. Result ID: 6a15624dfbc373592c0cff59. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reference":"OpenAI 2026"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.