Skip to content

Benchmark

SciCode

Coding benchmarks compare models on programming tasks. Scores are percentages. Higher is better.

Models measured
154
Scores recorded
265
Sources
3
IQ weight
1.71

Score distribution

47.7%
Mean
50.1%
Median
10.7%
Std dev
35.4%
10th percentile
58.8%
90th percentile
1.3%
Min
1.3%91.8%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.71 for chat models, measured over 154 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51GPT 5.4xhigh· Epoch AI Benchmarking Hub data56.6%
52GPT-6 Astramax· Epoch AI Benchmarking Hub data56.5%
53Grok 4.6high· Epoch AI Benchmarking Hub data56.5%
54Claude Fable 5.1medium· Epoch AI Benchmarking Hub data56.4%
55Claude Opus 5max· Epoch AI Benchmarking Hub data56.4%
56GPT-5.6 Sollow· Epoch AI Benchmarking Hub data56.4%
57GPT 5.5high· Epoch AI Benchmarking Hub data56.1%
58DeepSeek V4.1 Flash 1017max· Epoch AI Benchmarking Hub data55.9%
59Grok 4.6medium· Epoch AI Benchmarking Hub data55.9%
60GPT 5.5xhigh· Epoch AI Benchmarking Hub data55.8%
61GPT-6.1 Solhigh· Epoch AI Benchmarking Hub data55.8%
62GPT-6 Astraxhigh· Epoch AI Benchmarking Hub data55.7%
63GPT-6.1 Solxhigh· Epoch AI Benchmarking Hub data55.7%
64GPT-6 Astrahigh· Epoch AI Benchmarking Hub data55.4%
65Gemini 3.8 Flashmedium· Epoch AI Benchmarking Hub data55.1%
66GPT-6 Solxhigh· Epoch AI Benchmarking Hub data55.1%
67GPT-6.1 Terramax· Epoch AI Benchmarking Hub data55.1%
68Claude Haiku 5.5max· Epoch AI Benchmarking Hub data55%
69Claude Opus 5xhigh· Epoch AI Benchmarking Hub data55%
70Gemini 3.8 Flashlow· Epoch AI Benchmarking Hub data55%
71GPT-5.6 Terramax· Epoch AI Benchmarking Hub data55%
72Grok 4.5high· Epoch AI Benchmarking Hub data55%
73GPT-6 Solhigh· Epoch AI Benchmarking Hub data54.9%
74Grok 4.7low· Epoch AI Benchmarking Hub data54.9%
75GPT-6 Lunamax· Epoch AI Benchmarking Hub data54.6%
76Claude Opus 4.7max· Epoch AI Benchmarking Hub data54.5%
77GPT 5.5medium· Epoch AI Benchmarking Hub data54.5%
78Claude Opus 5high· Epoch AI Benchmarking Hub data54.3%
79Claude Sonnet 5high· Epoch AI Benchmarking Hub data54.3%
80GPT-6 Astramedium· Epoch AI Benchmarking Hub data54.2%
81GPT-6.1 Solmax· Epoch AI Benchmarking Hub data54.2%
82Claude Sonnet 5xhigh· Epoch AI Benchmarking Hub data54.1%
83GPT-6 Astralow· Epoch AI Benchmarking Hub data54.1%
84Gemini 3.5 Flashhigh· Epoch AI Benchmarking Hub data53.9%
85GPT-6 Solmedium· Epoch AI Benchmarking Hub data53.8%
86Claude Sonnet 5.5high· Epoch AI Benchmarking Hub data53.7%
87Claude Sonnet 5max· Epoch AI Benchmarking Hub data53.6%
88Gemini 3.7 Flashlow· Epoch AI Benchmarking Hub data53.6%
89GPT-5.6 Lunamax· Epoch AI Benchmarking Hub data53.6%
90Qwen3.7 Max· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
53.5%
91Qwen3.7 MaxQwen3.7-Max release · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.53.5%
92Claude Opus 4.8max· Epoch AI Benchmarking Hub data53.5%
93GPT-6 Astranone· Epoch AI Benchmarking Hub data53.5%
94Kimi K2.6· Epoch AI Benchmarking Hub data
Values from 2 sources
53.5%
95Gemini 3.6 Flashhigh· Epoch AI Benchmarking Hub data53.4%
96GPT-6.1 Terraxhigh· Epoch AI Benchmarking Hub data53.4%
97GPT-6.1 Sollow· Epoch AI Benchmarking Hub data53.2%
98GPT-6.1 Solmedium· Epoch AI Benchmarking Hub data53.2%
99Qwen3.8 Max· Epoch AI Benchmarking Hub data53.2%
100Grok 4.6xhigh· Epoch AI Benchmarking Hub data53%

Showing 51–100 of 265 results.

Data credits: Data by LLM Stats · Epoch AI Benchmarking Hub, CC BY 4.0 (licence) · BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.