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
201Mistral Medium2604· Epoch AI Benchmarking Hub data40.2%
202Claude Sonnet 4· Epoch AI Benchmarking Hub data40%
203Gemma 4 26B A4B· Epoch AI Benchmarking Hub data40%
204Qwen 3.8 27Blow· Epoch AI Benchmarking Hub data40%
205Step 3.7 Flash· Epoch AI Benchmarking Hub data40%
206Gemma 4 26B A4BGemma 4 Technical Report · Table 5 · thinking mode unless noted· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.40%
207DeepSeek V4 Pro 0813none· Epoch AI Benchmarking Hub data39.9%
208GPT-5 mini· Epoch AI Benchmarking Hub data39.2%
209Magistral Medium· Epoch AI Benchmarking Hub data39.2%
210DeepSeek V3· Epoch AI Benchmarking Hub data39%
211GPT-5 minihigh· Epoch AI Benchmarking Hub data39%
212Qwen 3.8 27Bmedium· Epoch AI Benchmarking Hub data39%
213MiniMax M2.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim39%
214DeepSeek V3.2 Expthinking· Epoch AI Benchmarking Hub data38.9%
215GPT OSS 20bhigh· Epoch AI Benchmarking Hub data38.9%
216Granite 4.2 30B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim38.8%
217Mercury 2· Epoch AI Benchmarking Hub data
Values from 2 sources
38.7%
218Command A Plus 05 2026· Epoch AI Benchmarking Hub data
Values from 2 sources
38.5%
219DeepSeek V4.1 Flash 1017none· Epoch AI Benchmarking Hub data38.5%
220Mercury 2.5· Epoch AI Benchmarking Hub data38.5%
221GLM 4.6· Epoch AI Benchmarking Hub data38.4%
222DeepSeek R1· Epoch AI Benchmarking Hub data38.3%
223North Mini Code 1.0· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim38.2%
224DeepSeek V3.1 Terminus· Epoch AI Benchmarking Hub data38%
225GLM 4.5 Air· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim37.3%
226Qwen3.6 27Bnone· Epoch AI Benchmarking Hub data37.3%
227Mistral Large 3· Epoch AI Benchmarking Hub data36.6%
228Qwen3.6 35B A3B· Epoch AI Benchmarking Hub data36.6%
229Qwen 3.8 27Bnone· Epoch AI Benchmarking Hub data36.2%
230Qwen3 Coder Next· Epoch AI Benchmarking Hub data36.2%
231GLM 5.2none· Epoch AI Benchmarking Hub data36.1%
232Granite 4.2 8B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim36.1%
233MiniMax M2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.36%
234GPT OSS 120Blow· Epoch AI Benchmarking Hub data36%
235Nemotron 3 Super· Epoch AI Benchmarking Hub data36%
236Qwen3.5 122B-A10Bnone· Epoch AI Benchmarking Hub data35.6%
237DeepSeek V4.1 Flashnone· Epoch AI Benchmarking Hub data35.5%
238Qwen3 32B· Epoch AI Benchmarking Hub data35.4%
239Magistral Small 2509· Epoch AI Benchmarking Hub data35.2%
240GPT OSS 120Bhigh· Epoch AI Benchmarking Hub data34%
241Nemotron 3 Nano 30B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim33.3%
242Qwen3 30B A3B2507, thinking· Epoch AI Benchmarking Hub data33%
243Nemotron 3.5 Lightning 30B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim32.6%
244Nemotron 3.5 Lightning 30B A3Bofficial release table· Reported by NVIDIA· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.32.6%
245Mistral Medium 3.1· Epoch AI Benchmarking Hub data32.4%
246Llama 4 Maverick· Epoch AI Benchmarking Hub data31.7%
247Qwen3 14B· Epoch AI Benchmarking Hub data31.6%
248Qwen3.5 9B· Epoch AI Benchmarking Hub data29.5%
249Qwen3.5 35B A3Bnone· Epoch AI Benchmarking Hub data29.3%
250Devstral Small 2505· Epoch AI Benchmarking Hub data28.8%

Showing 201–250 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.