SShogo Bench

Humanity's Last Exam

2,500 expert-written questions at the edge of human knowledge.

Agent Index weight
0%
shown for context only
Models tested
43
from the benchmark’s leaderboard and reports
Normalised for the index
0 → 0, 1 → 100
Official site

Results

#ModelScoreNormalisedHarnessCost/taskDate
1GPT-6 Astra Supported
OpenAI
54.8%source——
2Claude Fable 5.1 Supported
Anthropic
46.5%source——
3Gemini 3.1 Pro Preview Supported
Google
46.4%source——
4Claude Sonnet 5.5 Estimated
Anthropic
46.1%source——
5Gemini 3.8 Flash Supported
Google
44.5%source——
6GPT-5.4 Pro
OpenAI
44.3%source——
7Muse Spark
Meta
40.6%source——
8Gemini 3 Pro Preview Supported
Google
37.5%source——
9GPT-5.4 Supported
OpenAI
36.2%source——
10Claude Opus 4.7 Supported
Anthropic
36.2%source——
11Claude Opus 4.6 Supported
Anthropic
34.4%source——
12GPT-5 Pro
OpenAI
31.6%source——
13Claude Haiku 5.5 Estimated
Anthropic
30.1%source——
14GPT-5.2 Supported
OpenAI
27.8%source——
15GPT-5 Supported
OpenAI
25.3%source——
16Claude Opus 4.5 (latest) Supported
Anthropic
25.2%source——
17Kimi K2.5 Supported
Moonshot AI
24.4%source——
18GPT-5.1 Supported
OpenAI
23.7%source——
19Gemini 2.5 Pro Supported
Google
21.6%source——
20o3 Supported
OpenAI
20.3%source——
21GPT-5 Mini Estimated
OpenAI
19.4%source——
22o4-mini Estimated
OpenAI
18.1%source——
23Claude Sonnet 4.5 (latest) Supported
Anthropic
13.7%source——
24Gemini 2.5 Flash Estimated
Google
12.1%source——
25Claude Opus 4.1 (latest) Supported
Anthropic
11.5%source——
26Claude-Opus-4 Supported
Anthropic
10.7%source——
27Gemini 3.1 Flash Lite
Google
8.6%source——
28GLM-4.5
Z.ai (Zhipu)
8.3%source——
29GLM-4.5-Air
Z.ai (Zhipu)
8.1%source——
30o1-pro
OpenAI
8.1%source——
31Claude 3.7 Sonnet Supported
Anthropic
8.0%source——
32o1 Estimated
OpenAI
8.0%source——
33Claude Sonnet 4 Supported
Anthropic
7.8%source——
34Gemini 2.0 Flash Thinking (Jan 2025)
Google
6.6%source——
35Llama 4 Maverick
Meta
5.7%source——
36GPT-4.5
OpenAI
5.4%source——
37GPT-4.1
OpenAI
5.4%source——
38Gemini 1.5 Pro (Feb 2024)
Google
4.6%source——
39Mistral Medium 3
Mistral
4.5%source——
40Amazon Nova Pro
Amazon
4.4%source——
41Claude-Sonnet-3.5 Estimated
Anthropic
4.1%source——
42Amazon Nova Lite
Amazon
3.6%source——
43GPT-4o (2024-08-06) Supported
OpenAI
2.7%source——