Home Model series GPT unsuffixed models GPT-5.4 Humanity's Last Exam GPT-5.4 · Humanity's Last Exam (2,500-question benchmark) · Main text evidence (ev-gpt54-hle-tools) GPT-5.4 · Humanity's Last Exam (2,500-question benchmark) · Main text evidence (ev-gpt54-hle-tools) Official source for Humanity's Last Exam · 2,500-question benchmark in GPT-5.4. See “GPT-5.4 震撼登场 | OpenAI” at Evaluation / Academic / Humanity's Last Exam with tools row. Evidence ev-gpt54-hle-tools.
Source excerpt and location GPT-5.4 · Humanity's Last Exam (2,500-question benchmark), configuration GPT-5.4 (xhigh), Accuracy 52.1% (exact score publication date unconfirmed). See “GPT-5.4 震撼登场 | OpenAI” at Evaluation / Academic / Humanity's Last Exam with tools row. View score conditions
GPT-5.4 · Humanity's Last Exam (2,500-question benchmark), configuration GPT-5.4 Pro, Accuracy 58.7% (exact score publication date unconfirmed). See “GPT-5.4 震撼登场 | OpenAI” at Evaluation / Academic / Humanity's Last Exam with tools row. View score conditions
Humanity's Last Exam (with tools): GPT-5.4 52.1%; GPT-5.4 Pro 58.7%. Official material GPT-5.4 震撼登场 | OpenAI Location Main text · Evaluation / Academic / Humanity's Last Exam with tools row Material version Official article; page date label 2026-03-05; current page revision date unknown. Version published Unconfirmed Disclosure date Unconfirmed · Launch or later disclosure unconfirmed Collected 2026-11-11 Historical basis No reliable historical version evidence available Citation text GPT-5.4 · Humanity's Last Exam (2,500-question benchmark), configuration GPT-5.4 (xhigh), Accuracy 52.1% (exact score publication date unconfirmed); GPT-5.4 · Humanity's Last Exam (2,500-question benchmark), configuration GPT-5.4 Pro, Accuracy 58.7% (exact score publication date unconfirmed). Official source: GPT-5.4 震撼登场 | OpenAI (Main text, Official article; page date label 2026-03-05; current page revision date unknown.), Evaluation / Academic / Humanity's Last Exam with tools row, https://openai.com/zh-Hans-CN/index/introducing-gpt-5-4/. Benchlog https://benchlog.lekshome.cc/evidence/gpt-5-4/ev-gpt54-hle-tools/?lang=en
Related scores and conditions GPT-5.4 · Humanity's Last Exam scores and configurations Metric and score Configuration and evaluation conditions Evidence relation Accuracy 52.1% Reported 52.1% · Unit % Published Unconfirmed
GPT-5.4 (xhigh) Settings The release article states evaluations use xhigh unless otherwise noted. Unknowns none recorded Tools Tools allowed Reasoning budget xhigh Framework not recorded Protocol Humanity's Last Exam; with tools. Conditions Official release table labels GPT-5.4 (xhigh); the article says evaluations use xhigh unless otherwise noted and were run in a research environment. Directly supports this score
View this score in rankings Accuracy 58.7% Reported 58.7% · Unit % Published Unconfirmed
GPT-5.4 Pro Settings Pro model variant; the article reports Pro separately from GPT-5.4. Unknowns The Pro column is not a reasoning-effort label; the exact effort for each Pro evaluation is not separately stated. Tools Tools allowed Reasoning budget not recorded Framework not recorded Protocol Humanity's Last Exam; with tools. Conditions Official release table labels GPT-5.4 Pro; research evaluation environment. Directly supports this score
View this score in rankings