Skip to content

Benchmark

Arena ELO Webdev

Coding benchmarks compare models on programming tasks. Higher scores are better.

Models measured
139
Scores recorded
159
Sources
1
IQ weight
1.96

Score distribution

1,484.4
Mean
1,484.4
Median
157.2
Std dev
1,263.4
10th percentile
1,683
90th percentile
1,072
Min
1,0721,823.6
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.96 for chat models, measured over 139 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
101DeepSeek-V4-Flash PreviewHigh· LMArena data1,431.1
102GPT 5.2· LMArena data1,419
103GPT 5Medium· LMArena data1,417.5
104Inkling FP4· LMArena data1,414.1
105Inkling Small· LMArena data1,407.7
106GPT-5.3 CodexCodex Harness· LMArena data1,406.1
107Kimi K2.5Instant· LMArena data1,405.6
108NVIDIA Nemotron 3.5 Super VL 120BNVFP4· LMArena data1,400.7
109GPT 5.4· LMArena data1,400
110Qwen3.5 397B A17B· LMArena data1,399.9
111GPT-5.4 miniHigh· LMArena data1,398.2
112MiniMax M2.7· LMArena data1,398
113GPT 5.1Medium· LMArena data1,397.2
114Claude Sonnet 4.5High, 32K, 20250929· LMArena data1,393.9
115Claude Opus 4.120250805· LMArena data1,390.4
116MiniMax M2.5· LMArena data1,385.9
117Claude Sonnet 4.520250929· LMArena data1,384.8
118Gemini 3 FlashMinimal· LMArena data1,384.5
119MiniMax-M2.1 Preview· LMArena data1,384.2
120Grok 4.20Reasoning, 0309, beta· LMArena data1,376.4
121Gemma 4 31B IT· LMArena data1,370.2
122Solar Pro 4· LMArena data1,369.2
123DeepSeek V3.2Thinking· LMArena data1,362.5
124Qwen3.5 122B-A10B· LMArena data1,360.2
125Gemma 4 26B A4B· LMArena data1,357.6
126Grok 4.3· LMArena data1,357.5
127Qwen3.5 27B· LMArena data1,357
128Muse Glimmer· LMArena data1,353
129Laguna M.1· LMArena data1,350.1
130GPT 5.1· LMArena data1,342.3
131GLM 4.6· LMArena data1,342
132GPT-5.2 Codex· LMArena data1,338.7
133GPT-5.1 Codex· LMArena data1,338
134MiMo V2 FlashNon-Thinking· LMArena data1,331.6
135Claude Haiku 4.520251001· LMArena data1,330.2
136DeepSeek V3.2· LMArena data1,325.7
137Kimi K2 Thinking· LMArena data1,321.8
138Laguna XS.2· LMArena data1,304
139MiniMax M2· LMArena data1,298.9
140MiMo V2 FlashThinking· LMArena data1,291
141Qwen3 Coder· LMArena data1,274.4
142DeepSeek V3.2 Exp· LMArena data1,271.4
143Mistral Medium 3.5· LMArena data1,263.4
144KAT-Coder-Pro V1· LMArena data1,257.4
145Gemini 3.1 Flash Lite· LMArena data1,255.8
146Qwen3.5 35B A3B· LMArena data1,253.4
147GPT-5.1 Codex Mini· LMArena data1,243.5
148Qwen3.5 Flash· LMArena data1,243.5
149Grok 4.1 FastReasoning· LMArena data1,241.7
150Trinity Large Thinking· LMArena data1,236.6

Showing 101–150 of 159 results.

Data credits: LMArena, CC BY 4.0 (licence)