PhysicianBench
We show this table for reference; we do not rank on it.
A public benchmark of 100 physician tasks carried out inside an electronic health record system.
Benchmark score on PhysicianBench — November 11, 2026
We compile the PhysicianBench rows from provider self-reports. Claude Opus 5.5 leads the table at 68.4%, followed by Claude Fable 5.2 (64.4%) and Claude Sonnet 5.5 (63.2%). We do not use these results to rank models overall.
Claude Opus 5.5
Anthropic
Claude Fable 5.2
Anthropic
Claude Sonnet 5.5
Anthropic
5 modelsKnowledgeCurrentDisplay onlyUpdated November 11, 2026
Benchmark score table (5 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | Claude Opus 5.5Anthropic | 68.4% | Not reported | Closed |
| 2 | Claude Fable 5.2Anthropic | 64.4% | Not reported | Closed |
| 3 | Claude Sonnet 5.5Anthropic | 63.2% | Not reported | Closed |
| 4 | Claude Fable 5.1Anthropic | 61.0% | Not reported | Closed |
| 5 | GPT-6 AstraOpenAI | 48.0% | Not reported | Closed |
Among the reported PhysicianBench rows, Claude Opus 5.5 is first at 68.4%. The third row is 5.2 points behind. The broader top-10 range is 20.4 points, so the table still separates the published systems.
5 models have been evaluated on PhysicianBench. The benchmark falls in the Knowledge category. PhysicianBench is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About PhysicianBench
Year
2026
Tasks
100 EHR physician tasks
Format
Pass rate
Difficulty
Clinical agent work
Anthropic reports the share of attempts passed over 100 tasks with five attempts each (one for GPT-6 Astra), at max effort. BenchLM first added PhysicianBench from the Claude Fable 5.2 system card and keeps it display-only.
Freshness and provenance
Version
PhysicianBench 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does PhysicianBench measure?
A public benchmark of 100 physician tasks carried out inside an electronic health record system.
Which model scores highest on PhysicianBench?
Claude Opus 5.5 by Anthropic currently leads with a score of 68.4% on PhysicianBench.
How many models are evaluated on PhysicianBench?
5 AI models have published results on PhysicianBench in the BenchLM catalog.
Compare top models on PhysicianBench
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.