FHIR-AgentBench (Multi-Turn + Retriever + Code): leaderboard
Metric: Answer correctness (%; o4-mini judge of the free-text answer against the SQL-derived answer, 2,931 questions; multi-turn agent with a retriever and a code interpreter). Source: arxiv.org. Saturation forecast: Around April 2028. 4 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O4 Mini | 50 |
| 2 | Qwen 3 32B | 47 |
| 3 | Llama 3.3 70B | 46 |
| 4 | Gemini 2.5 Flash | 44 |
Interactive version: theaggregate.ai/benchmark?slug=fhir-agentbench-multi-turn-plus-retriever-plus-code · How It Works · Data refreshed daily, snapshot 2026-09-26.