Home Model series Gemini 4 Carbon Gemini 4 Carbon Gemini 4 Carbon · Agents' Last Exam · v1.1, 122 tasks released 2026-10-17 (before 2026-10-26 update) · Scores and configurations Gemini 4 Carbon · Agents' Last Exam · v1.1, 122 tasks released 2026-10-17 (before 2026-10-26 update) · Scores and configurations Gemini 4 Carbon record for Agents' Last Exam · v1.1, 122 tasks released 2026-10-17 (before 2026-10-26 update). 1 score record for this release, including 1 exact value (may include historical records). Includes 1 source location.
Scores and evaluation conditions Gemini 4 Carbon · Agents' Last Exam scores and configurations Metric and score Configuration and evaluation conditions Sources Average rubric score 77.4% Reported 77.4% · Unit % Published Unconfirmed
Gemini 4 Carbon highest thinking Settings Gemini API model gemini-4-carbon; highest thinking settings; default sampling unless benchmark method states otherwise; pass@1 except where stated. Unknowns Exact thinking budget and sampling parameters are not published.、原文仅称 highest thinking settings,未确认对应的公开档位名称。 Tools Tool access not recorded Reasoning budget not recorded Framework ALE-Claw Protocol 122-task v1.1 set released 2026-10-17, default ALE-Claw, 5-hour window, safety filters enabled; average rubric score. Conditions Uses task version before the 2026-10-26 update; Gemini responses flagged by safety filters become empty strings while episode may continue. View this score in rankings
Benchmark versions and metrics Aliases Agents' Last Exam v1.1
Average rubric score · % · Higher is better