Benchmark

Agents' Last Exam

Agents' Last Exam is a challenging benchmark for AI agents on hard, long-horizon tasks that test sustained reasoning, planning, and tool use, reported…

Modality
text
Categories
reasoning, agents, tool_calling
Openness
unknown
Source
llm_stats
Reported scores
10

Reported scores

llm_stats

ModelOrganizationReported valueReportedEvidence
DeepSeek-V4-Flash-0731DeepSeek0.2522026-07-31evidence
DeepSeek-V4-Pro-0813DeepSeek0.2572026-08-13evidence
Gemini 3.7 FlashGoogle0.2632026-08-13evidence
GLM-5.3Z.ai0.2852026-08-14evidence
GPT-5.6 LunaOpenAI0.5032026-07-09evidence
GPT-5.6 SolOpenAI0.5272026-07-09evidence
GPT-5.6 TerraOpenAI0.5042026-07-09evidence
Qwen3.8 MaxQwen0.5242026-08-02evidence
Qwen3.8-27BQwen0.4292026-08-14evidence
Seed 2.1 ProByteDance0.4142026-06-24evidence

Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.