Skip to content
Trending storyDeveloping

Claude Sonnet 5.5 and GPT-6 Astra tie at 41.7% on MLCR-AA medical benchmark

1 article1 sourcesince Oct 21Last article 1h ago ·

Overview

AISummary of 1 article

Artificial Analysis updated its MLCR-AA leaderboard, which tests long-context medical reasoning on Wisedocs' benchmark, and lists Claude Sonnet 5.5 (max with fallback) and GPT-6 Astra (max) tied at the top with 41.7%.

GPT-6.1 Sol (max) is third at 38.3%. Artificial Analysis says answers are now graded by a three-model judge panel, and that it reviewed all 60 expert reference answers against their source records, refining 15 of them.

According to the leaderboard post, no model passes more than 42% of questions. Claude models lead on completeness and OpenAI models lead on accuracy.

Written by AI from the articles below · updated Oct 21, 12:09 PM ET

Check the sources:

Latest developmentOct 21, 11:53 AM ET
Artificial Analysis updates MLCR-AA, with Claude Sonnet 5.5 and GPT-6 Astra tied at 41.7%

Article timeline

The articles in this story. Times are ET.

Oct 21
  1. Artificial AnalysisOfficial
    Artificial Analysis updates MLCR-AA, with Claude Sonnet 5.5 and GPT-6 Astra tied at 41.7%

    AIArtificial Analysis updated its MLCR-AA leaderboard for Wisedocs' medical long-context reasoning benchmark, with models now prompted as medical record reviewers and graded by a three-model judge panel. Claude Sonnet 5.5 (max with fallback) and GPT-6 Astra (max) are tied at the top with 41.7%, ahead of GPT-6.1 Sol (max) at 38.3%. No model passes more than 42% of questions, and Claude models lead on completeness while OpenAI models lead on accuracy.

    Image from @ArtificialAnlys's post

Heat Trend

Not enough continuous observations to show a trend yet.