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
101Claude Sonnet 5.5medium· Epoch AI Benchmarking Hub data52.9%
102Kimi K3low· Epoch AI Benchmarking Hub data52.7%
103GPT-5.5 Instant· Epoch AI Benchmarking Hub data52.5%
104GPT-5.6 Terrahigh· Epoch AI Benchmarking Hub data52.4%
105GPT-6.1 Terralow· Epoch AI Benchmarking Hub data52.4%
106GPT-5.6 Terraxhigh· Epoch AI Benchmarking Hub data52.3%
107Inkling 2 Small· Epoch AI Benchmarking Hub data52.3%
108Kimi K2.6Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · avg@10· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.52.2%
109GPT-5.4 minixhigh· Epoch AI Benchmarking Hub data52.1%
110DeepSeek V4.1 Flashmax· Epoch AI Benchmarking Hub data51.9%
111Claude Haiku 5.5xhigh· Epoch AI Benchmarking Hub data51.7%
112DeepSeek V4 Pro 0813high· Epoch AI Benchmarking Hub data51.7%
113GPT-6 Lunaxhigh· Epoch AI Benchmarking Hub data51.7%
114Claude Sonnet 5medium· Epoch AI Benchmarking Hub data51.6%
115GLM 5.3 Flash· Epoch AI Benchmarking Hub data51.6%
116GPT 5.5low· Epoch AI Benchmarking Hub data51.6%
117GPT-5.6 Lunahigh· Epoch AI Benchmarking Hub data51.6%
118GPT-6.1 Terrahigh· Epoch AI Benchmarking Hub data51.6%
119GPT-6.1 Terramedium· Epoch AI Benchmarking Hub data51.6%
120Muse Spark· Epoch AI Benchmarking Hub data51.5%
121Qwen3.7 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
51.3%
122Qwen3.7 PlusQwen3.7-Plus release · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.51.3%
123MiMo V2.6 Flash· Epoch AI Benchmarking Hub data51.3%
124GLM 5.2max· Epoch AI Benchmarking Hub data51.2%
125DeepSeek V4 Pro 0813max· Epoch AI Benchmarking Hub data51%
126GPT-6 Lunamedium· Epoch AI Benchmarking Hub data50.9%
127Claude Opus 5medium· Epoch AI Benchmarking Hub data50.7%
128GPT-5.6 Lunaxhigh· Epoch AI Benchmarking Hub data50.5%
129GPT-5.6 Terramedium· Epoch AI Benchmarking Hub data50.5%
130GPT-6 Lunahigh· Epoch AI Benchmarking Hub data50.3%
131GPT-6 Sollow· Epoch AI Benchmarking Hub data50.2%
132Grok Build 0.1· Epoch AI Benchmarking Hub data50.2%
133MiMo V2.5 Pro· Epoch AI Benchmarking Hub data50.2%
134Claude Sonnet 5low· Epoch AI Benchmarking Hub data50.1%
135MiniMax M2.7· Epoch AI Benchmarking Hub data50.1%
136DeepSeek V4 Promax· Epoch AI Benchmarking Hub data50%
137DeepSeek V4 Flash 0731max· Epoch AI Benchmarking Hub data49.9%
138GPT-5.6 Terralow· Epoch AI Benchmarking Hub data49.9%
139Inkling Small· Epoch AI Benchmarking Hub data
Values from 2 sources
49.7%
140Grok 4.6low· Epoch AI Benchmarking Hub data49.4%
141Beam 1.0 501B A23B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim49.2%
142Claude Haiku 5.5low· Epoch AI Benchmarking Hub data49.2%
143Claude Sonnet 5.5low· Epoch AI Benchmarking Hub data49.1%
144Claude Haiku 5.5medium· Epoch AI Benchmarking Hub data49%
145Kimi K2.5· Epoch AI Benchmarking Hub data
Values from 2 sources
49%
146Claude Haiku 5.5high· Epoch AI Benchmarking Hub data48.7%
147Kimi K2.5setting not named by the source· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.48.7%
148Grok 4.3high· Epoch AI Benchmarking Hub data48.3%
149Claude Opus 5low· Epoch AI Benchmarking Hub data48%
150GPT-6.1 Terranone· Epoch AI Benchmarking Hub data47.8%

Showing 101–150 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.