Anthropic's Claude Sonnet 5.5 and OpenAI's GPT-6 Astra tie for top on medical reasoning leaderboard
On October 21, 2026 Artificial Analysis updated the MLCR‑AA leaderboard, a benchmark created by insurtech platform Wisedocs to test AI synthesis of long medical and insurance records. The update shows Anthropic’s Claude Sonnet 5.5 and OpenAI’s GPT‑6 Astra tied for first place with a 41.7% pass rate.
Key points
- Claude Sonnet 5.5 and GPT-6 Astra each achieved a 41.7% pass rate, tying for first on the MLCR-AA leaderboard (Oct 21 update).
- Median score across all models remains below 15%; completeness leader Claude Sonnet 5.5 scored 64.8%, accuracy leader GPT-6.1 Sol scored 96.7%.
- Benchmark uses synthetic medical/insurance records of 25,000‑64,000 tokens to assess AI’s ability to synthesize data for claims professionals.
The median score across all 60 models stays below 15%, and no model exceeds 42%. Claude Sonnet 5.5 leads the completeness metric at 64.8%, while OpenAI’s GPT‑6.1 Sol tops accuracy at 96.7%. The benchmark feeds synthetic case files of 25,000‑64,000 tokens, allowing realistic testing without real patient data. The October update also added practical reviewer prompts and refined 15 of the 60 expert references.
For insurers, the results highlight a trade‑off: Anthropic’s models capture more information, whereas OpenAI’s reduce confident errors. Even the best models still fail the majority of expert‑level synthesis tasks, underscoring the current limits of AI in long‑context claims work.
Model pages: Claude Sonnet 5.5 → · GPT-6 Astra → · GPT-6.1 Sol →
The story so far
2 episodes →- Anthropic's Claude Sonnet 5.5 and OpenAI's GPT-6 Astra tie for top on medical reasoning leaderboardthis story
Claude Sonnet 5.5 and GPT-6 Astra tie for top spot on MLCR-AA medical reasoning leaderboard
cryptobriefing.com · 21 October 2026
Loading the full article…
This text was published by cryptobriefing.com and written by Diego Almada Lopez. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Generative AI & Models
All →- Nvidia releases Kumo Forecast time-series model with retrieval and interpretability · 3 src
- OpenAI's GPT-6 Astra cracks Napoleon cipher; Anthropic's Claude maps NSA elliptic seeds · 1 src
- Dharma AI says VocabX reduces inference cost by 77% for Qwen3-4B grader · 1 src
- Recursive releases Flint 40B coding agent model on Hugging Face · 3 src
- Anthropic cuts Haiku 5.5 API costs by 90% for short inputs · 4 src
Comments
via GitHub Discussions