Benchgen

Agents' Last Exam — Results

RankModelScore
1gpt-6-astra59.3
2gpt-6-sol56.4
3deepseek-v4-1-flash31.8
4deepseek-v4-1-flash31.8
5deepseek-v4-flash-vision-exp27.3
6gpt-5-6-sol0.527
7qwen3-8-max0.524
8qwen3-8-flash-next0.512
9gpt-5-6-terra0.504
10gpt-5-6-luna0.503
11claude-fable-50.487
12gpt-5-50.479
13claude-opus-4-80.451
14qwen3-8-27b0.429
15claude-opus-4-70.418
16seed-2-1-pro0.414
17glm-5-20.406
18gpt-5-40.373
19gemini-3-1-pro0.32
20qwen3-7-max0.311
21kimi-k30.283
22glm-5-3-flash0.263
23deepseek-v4-pro0.257
24qwen3-6-plus0.243
25kimi-k2-60.217
A

Agents' Last Exam — Leaderboard

1 phaseActive

Agents' Last Exam Leaderboard

No self-reported entries.