Home Model series Claude Opus Claude Opus 5 Claude Opus 5 · HealthBench · 48,000+ rubric items across 5,000 multi-turn conversations · Scores and configurations Claude Opus 5 · HealthBench · 48,000+ rubric items across 5,000 multi-turn conversations · Scores and configurations Claude Opus 5 record for HealthBench · 48,000+ rubric items across 5,000 multi-turn conversations. 2 score records for this release, including 2 exact values (may include historical records). Includes 1 source location.
Scores and evaluation conditions Claude Opus 5 · HealthBench scores and configurations Metric and score Configuration and evaluation conditions Sources Raw score 67.1% Reported 67.1% · Unit % Published Unconfirmed
HealthBench · max Settings Max effort; no tools or customized system prompts; Opus 4.8 grader; five-trial average; length-adjustment method cited from the GPT-5.5 system card. Unknowns none recorded Tools Tools forbidden Reasoning budget max Framework not recorded Protocol HealthBench; raw rubric score. Conditions Max effort; no tools or customized system prompt; Opus 4.8 grader; five-trial average. View this score in rankings Length-adjusted score 57.8% Reported 57.8% · Unit % Published Unconfirmed
HealthBench · max Settings Max effort; no tools or customized system prompts; Opus 4.8 grader; five-trial average; length-adjustment method cited from the GPT-5.5 system card. Unknowns none recorded Tools Tools forbidden Reasoning budget max Framework not recorded Protocol HealthBench; length-adjusted score. Conditions Max effort; no tools or customized system prompt; Opus 4.8 grader; five-trial average. View this score in rankings
Benchmark versions and metrics Aliases HealthBench
Length-adjusted score · % · Higher is better Raw score · % · Higher is better Mean response length · characters · Ranking direction unconfirmed 长度校正分(0–100) · score · Higher is better 未校正分(0–100) · score · Higher is better 相对 GPT-5.4 的长度校正分差 · points · Ranking direction unconfirmed Unadjusted score · % · Higher is better