Skip to content

Benchmark

HLE

Scores are percentages. Higher is better.

Models measured
48
Scores recorded
71
Sources
2
IQ weight
2.01

Score distribution

20.3%
Mean
0.253%
Median
31.1%
Std dev
0.054%
10th percentile
70%
90th percentile
0.0272%
Min
0.0272%77.4%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.01 for chat models, measured over 48 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, perplexity, server-tool, 25, high· OpenRouter benchmark exports77.4%
2Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, perplexity, server-tool, 1, high· OpenRouter benchmark exports75.3%
3Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, exa, server-tool, 25, high· OpenRouter benchmark exports74.2%
4Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, perplexity, server-tool, 5, high· OpenRouter benchmark exports71.7%
5Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, exa, server-tool, 5, high· OpenRouter benchmark exports71.4%
6GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, native, server-tool, 25, high· OpenRouter benchmark exports71.1%
7GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, perplexity, server-tool, 25, high· OpenRouter benchmark exports71%
8GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, perplexity, server-tool, 1, high· OpenRouter benchmark exports70%
9GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, perplexity, server-tool, 5, high· OpenRouter benchmark exports69%
10Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, parallel, server-tool, 25, high· OpenRouter benchmark exports68.6%
11Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, exa, server-tool, 1, high· OpenRouter benchmark exports68.5%
12GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, exa, server-tool, 25, high· OpenRouter benchmark exports68.4%
13GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, parallel, server-tool, 25, high· OpenRouter benchmark exports66.4%
14GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, native, server-tool, 1, high· OpenRouter benchmark exports66%
15GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, native, server-tool, 5, high· OpenRouter benchmark exports66%
16Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, parallel, server-tool, 5, high· OpenRouter benchmark exports65%
17GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, exa, server-tool, 5, high· OpenRouter benchmark exports65%
18GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, parallel, server-tool, 5, high· OpenRouter benchmark exports64.3%
19GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, exa, server-tool, 1, high· OpenRouter benchmark exports64%
20Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, parallel, server-tool, 1, high· OpenRouter benchmark exports60.1%
21GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, parallel, server-tool, 1, high· OpenRouter benchmark exports57.9%
22GPT-6 Astra· Epoch AI Benchmarking Hub data54.8%
23Claude Fable 5.1xhigh· Epoch AI Benchmarking Hub data46.5%
24Gemini 3.1 Pro Preview· Epoch AI Benchmarking Hub data46.4%
25Claude Sonnet 5.5xhigh· Epoch AI Benchmarking Hub data46.1%
26Gemini 3.8 Flash· Epoch AI Benchmarking Hub data44.5%
27GPT-5.4 Pro· Epoch AI Benchmarking Hub data44.3%
28Muse Spark· Epoch AI Benchmarking Hub data40.6%
29Gemini 3 Pro Preview· Epoch AI Benchmarking Hub data37.5%
30GPT 5.4xhigh· Epoch AI Benchmarking Hub data36.2%
31Claude Opus 4.7· Epoch AI Benchmarking Hub data36.2%
32Claude Opus 4.6max· Epoch AI Benchmarking Hub data34.4%
33GPT-5 Pro· Epoch AI Benchmarking Hub data31.6%
34Claude Haiku 5.5· Epoch AI Benchmarking Hub data30.1%
35GPT 5.2· Epoch AI Benchmarking Hub data27.8%
36GPT 5· Epoch AI Benchmarking Hub data25.3%
37Claude Opus 4.5· Epoch AI Benchmarking Hub data25.2%
38Kimi K2.5· Epoch AI Benchmarking Hub data24.4%
39GPT 5.1· Epoch AI Benchmarking Hub data23.7%
40Gemini 2.5 Pro· Epoch AI Benchmarking Hub data21.6%
41OpenAI o3high· Epoch AI Benchmarking Hub data20.3%
42GPT-5 mini· Epoch AI Benchmarking Hub data19.4%
43OpenAI o3medium· Epoch AI Benchmarking Hub data19.2%
44Claude Opus 4.6· Epoch AI Benchmarking Hub data19%
45Gemini 2.5 Pro· Epoch AI Benchmarking Hub data18.2%
46o4 Minihigh· Epoch AI Benchmarking Hub data18.1%
47Gemini 2.5 Pro Preview 05-06· Epoch AI Benchmarking Hub data17.8%
48o4 Minimedium· Epoch AI Benchmarking Hub data14.3%
49Claude Sonnet 4.5· Epoch AI Benchmarking Hub data13.7%
50gemini-2.5-flash-preview-04-17· Epoch AI Benchmarking Hub data12.1%

Showing 1–50 of 71 results.

Data credits: Epoch AI Benchmarking Hub, CC BY 4.0 (licence) · OpenRouter public benchmark leaderboard and search-lane exports, CC BY 4.0. Fractional source scores converted to percentages; immutable dated snapshots. (licence)