Home Model series GPT unsuffixed models GPT-5.4 GPT-5.4 · HealthBench · 48,000+ rubric items across 5,000 multi-turn conversations · Scores and configurations GPT-5.4 · HealthBench · 48,000+ rubric items across 5,000 multi-turn conversations · Scores and configurations GPT-5.4 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 2 source locations.
Scores and evaluation conditions GPT-5.4 · HealthBench scores and configurations Metric and score Configuration and evaluation conditions Sources Length-adjusted score 62.6% Reported 62.6% · Unit % Published Unconfirmed
GPT-5.4 system-card evaluation · effort unspecified Settings GPT-5.4 as named in system-card Table 6; reasoning effort and tool settings are not stated. Unknowns Reasoning effort、Tool access, grader details, and run settings Tools Tool access not recorded Reasoning budget not recorded Framework not recorded Protocol HealthBench; 5,000 realistic health conversations; example-specific rubrics. Table 6 reports one score without a separate raw/adjusted split. Conditions GPT-5.4, as named in the current saved system card; reasoning effort and evaluation run settings are not stated. Mean response length 3311 characters Reported 3311 chars · Unit characters Published Unconfirmed
GPT-5.4 system-card evaluation · effort unspecified Settings GPT-5.4 as named in system-card Table 6; reasoning effort and tool settings are not stated. Unknowns Reasoning effort、Tool access, grader details, and run settings Tools Tool access not recorded Reasoning budget not recorded Framework not recorded Protocol Mean response length on HealthBench, reported in characters. Conditions GPT-5.4, as named in the current saved system card; reasoning effort and evaluation run settings are not stated.
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