Skip to content

Benchmark

Vals LiveCodeBench

Coding benchmarks compare models on programming tasks. Scores are percentages.

Models measured
127
Scores recorded
134
Sources
1
IQ weight
2.89

Score distribution

71.5%
Mean
80.5%
Median
18.2%
Std dev
41.9%
10th percentile
87.5%
90th percentile
9.9%
Min
9.9%90.5%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.89 for chat models, measured over 127 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-5.6 Sol· Vals AI82.6%
52Claude Sonnet 5· Vals AI82.4%
53GLM 4.7· Vals AI82.2%
54o4 Mini· Vals AI82.2%
55MiniMax M3· Vals AI82.2%
56Claude Sonnet 4.6· Vals AI82.1%
57Kimi K2.7 Code· Vals AI82.1%
58GLM 5thinking· Vals AI81.9%
59MiniMax M2.1· Vals AI81.8%
60MiMo V2.5· Vals AI81.5%
61GPT-5.4 mini· Vals AI81.5%
62GLM 5.1· Vals AI81.4%
63MiMo V2.5 Pro· Vals AI81.4%
64GLM 4.6· Vals AI81%
65DeepSeek V3.2· Vals AI80.7%
66Grok 4.1 Fast· Vals AI80.6%
67GLM 5.3· Vals AI80.5%
68GLM 5.3 Flash· Vals AI80.5%
69GPT OSS 20b· Vals AI80.4%
70Gemini 3.1 Flash Lite· Vals AI80.1%
71MiniMax M2.7· Vals AI79.9%
72MiniMax M2.5· Vals AI79.2%
73Gemini 2.5 Pro Preview 03-25· Vals AI79.2%
74Gemini 3.5 Flash Lite· Vals AI79%
75Grok 4 Fast· Vals AI79%
76Qwen3 Max· Vals AI78.2%
77Grok 3 Mini Fast Betahigh, reasoning· Vals AI76.2%
78gemini-2.5-flash-preview-09-2025thinking· Vals AI76.2%
79gemini-2.5-flash-preview-09-2025· Vals AI75.1%
80Magistral Medium· Vals AI74.9%
81Claude Sonnet 4.5· Vals AI73%
82Magistral Small 2509· Vals AI72.1%
83o3 Mini· Vals AI71.5%
84Gemini 2.5 Flash Lite Preview 09-2025thinking· Vals AI71.4%
85Qwen 3 235B A22B· Vals AI70.6%
86Kimi K2 Instruct· Vals AI70.4%
87DeepSeek R1· Vals AI70.2%
88GPT-5 nano· Vals AI70.2%
89Claude Opus 4thinking· Vals AI70.2%
90GLM 5.2· Vals AI69.5%
91Laguna M.1· Vals AI68.1%
92Laguna XS.2· Vals AI67.8%
93Gemini 2.5 Flash Lite Preview 09-2025· Vals AI67.7%
94GLM 4.5· Vals AI67.4%
95Qwen3 Max Preview· Vals AI66.9%
96Claude Opus 4.1· Vals AI66.5%
97Grok 3 Mini Fast Betalow, reasoning· Vals AI66.3%
98DeepSeek V3· Vals AI65.5%
99Kimi K2 Thinking· Vals AI63.1%
100Claude Opus 4· Vals AI62.6%

Showing 51–100 of 134 results.