Frontier Log

Benchmark · Active

Agents' Last Exam v1.1

Created by UC Berkeley RDI, 14 Oct 2026 · Higher is better · Website

020406080Oct 202677.4%Gemini 4 Carbon69.0%Gemini 4 Argon020406080Oct 202677.4%Gemini 4 Carbon69.0%Gemini 4 Argon

Results

ModelScoreSourceSource check
Gemini 4 Carbon28 Oct 2026 · Google77.4%Gemini 4 Carbon: our next frontier model122 tasks; ALE-Claw harness; 5-hour window; October 17, 2026 task files; safety filters onSource image
Gemini 4 Argon30 Sep 2026 · Google69.0%Gemini 4 Carbon: our next frontier model122 tasks; ALE-Claw harness; 5-hour window; October 17, 2026 task files; safety filters on; reported laterSource image

All benchmarks