Frontier Log

Benchmark · InactiveSuperseded by Agents' Last Exam v1.1

Agents' Last Exam v1

Created by UC Berkeley RDI, 3 Jun 2026 · Higher is better · Website

01020304050Jul 2026Oct 202639.5%Gemini 4 Argon24.3%Qwen3.8-Flash-Next26.3%Gemini 3.7 Flash24.2%Gemini 3.6 Flash13.2%Qwen3.7-Plus01020304050Jul 2026Oct 202639.5%Gemini 4 Argon24.3%Qwen3.8-Flash-Next26.3%Gemini 3.7 Flash24.2%Gemini 3.6 Flash13.2%Qwen3.7-Plus

Settings and run setups

Task setScoringProviders
All tasks (152)Pass rateAlibabaMiniMaxZ.ai
CLI subset (105)Pass rateDeepSeekZ.ai
All tasks (152)ScoreOpenAI
CLI subset (105)ScoreOpenAI
Not statedPass rateGoogleAlibaba
Not statedScoreAlibaba
CLI subset (105)Not statedZ.aiTencent
Not statedNot statedOpenAIZ.ai

Download image

Results

ModelScoreSource
Gemini 4 Argon30 Sep 202639.5%Gemini 4 Argon: our next era of frontier intelligenceChecked A person checked this number against its source on 30 Oct 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.ALE-Claw harness · 5-hour window · safety filters onSource image
Gemini 3.7 Flash13 Aug 202626.3%Introducing Gemini 3.7 FlashCorroborated The same number is printed in 1 other lab document: 1ALE-Claw harnessSource image
Qwen3.8-Flash-Next26 Aug 202624.3%Qwen3.8-Flash-Next: A New Architecture, Towards Ultimate Cost-EfficiencyCorroborated The same number is printed in 1 other lab document: 1harness not stated · pass@1 · subset not statedOur capture
Gemini 3.6 Flash21 Jul 202624.2%Introducing Gemini 3.7 FlashCorroborated The same number is printed in 1 other lab document: 1ALE-Claw harness · reported laterSource image
Qwen3.7-Plus1 Jun 202613.2%Qwen3.8-Flash-Next: A New Architecture, Towards Ultimate Cost-EfficiencyCorroborated The same number is printed in 1 other lab document: 1harness not stated · pass@1 · subset not stated · reported laterOur capture

All benchmarks