Benchgen

Humanity's Last Exam — Results

RankModelScore
1claude-opus-5-567.7
2claude-sonnet-5-564.5
3deepseek-v4-1-flash63.9
4claude-fable-5-160.9
5claude-sonnet-557.4
6gpt-6-astra57.2
7ornith-1-5-397b56.1
8apodex-1-156.1
9kimi-k356
10seed-2-1-pro55.7
11glm-5-3-flash55.3
12glm-5-254.7
13seed-2-1-turbo54.6
14kimi-k2-thinking-090551
15fugu-ultra50
16qwen3-5-27b48.5
17deepseek-v4-pro-max48.2
18qwen3-5-122b-a10b47.5
19qwen3-5-35b-a3b47.4
20fugu47.2
21gemini-3-1-pro46.44
22inkling46
23deepseek-v4-flash-max45.1
24gpt-5-444.32
25qwen3-8-max43.6

Humanity's Last Exam — Evaluations

1 phaseActive

No evaluations yet for this benchmark.