Skip to content

Benchmark

Arena ELO - Industry Writing and Literature and Language

People compare model answers and choose which they prefer. Higher scores are better.

Models measured
365
Scores recorded
422
Sources
1
IQ weight
1.17

Score distribution

1,297.8
Mean
1,328.7
Median
142.8
Std dev
1,086
10th percentile
1,453.9
90th percentile
809.1
Min
809.11,540.8
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.17 for chat models, measured over 365 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
101ERNIE-5.0-Preview1022· LMArena data1,409.2
102DeepSeek V3.1 TerminusThinking· LMArena data1,409.2
103DeepSeek-V4-Flash PreviewHigh· LMArena data1,409
104Gemini 3.1 Flash Lite· LMArena data1,408.8
105MiMo V2.5· LMArena data1,408.3
106DeepSeek V3.1 Terminus· LMArena data1,407.8
107Qwen3 Max Preview· LMArena data1,407.5
108DeepSeek V3.2· LMArena data1,407.4
109Grok 4.1· LMArena data1,407.3
110DeepSeek V3.2 Exp· LMArena data1,407
111Nemotron 3 UltraNVFP4· LMArena data1,406.8
112MiniMax M3· LMArena data1,406.6
113Gemma 4 26B A4B· LMArena data1,406.5
114Gemini 2.5 Flash· LMArena data1,405.4
115Grok 4.1Thinking· LMArena data1,404.5
116DeepSeek V3.1Thinking· LMArena data1,404.4
117GLM 4.5· LMArena data1,402.4
118Qwen 3.8 27B· LMArena data1,401.7
119DeepSeek R10528· LMArena data1,401.4
120Grok 40709· LMArena data1,398.9
121DeepSeek V3.2Thinking· LMArena data1,398.7
122Kimi K2.5Instant· LMArena data1,397.7
123DeepSeek V3.1· LMArena data1,397.4
124Mistral Large 3· LMArena data1,397.3
125Beam 1.0 501B A23B· LMArena data1,397.1
126MiMo V2 Omni· LMArena data1,396.9
127DeepSeek V3.2 ExpThinking· LMArena data1,395.5
128Mistral Medium 3.12508· LMArena data1,395.3
129Gemini 2.5 Flash Preview09-2025· LMArena data1,394.8
130Grok 4.7xHigh· LMArena data1,394.8
131GPT-5.4 miniHigh· LMArena data1,392.4
132Claude Opus 4Thinking 16K, 20250514· LMArena data1,391.9
133GPT 5.2High· LMArena data1,391.7
134Qwen3 VL 235B· LMArena data1,391.6
135Grok 4 FastNon-Reasoning· LMArena data1,390.8
136GPT 5.2· LMArena data1,390.7
137Qwen3 Max2025-09-23· LMArena data1,390.1
138nova-experimental-chat26-02-10· LMArena data1,389.9
139Qwen3 235B A22B Instruct 25072507· LMArena data1,389.7
140GPT-6 SolMax· LMArena data1,389.6
141Hunyuan Vision 1.5 Thinking· LMArena data1,388.4
142Qwen 3 235B A22B2507, thinking· LMArena data1,387.8
143LongCat-Flash-Chat Exp2602· LMArena data1,387.6
144Grok 4.1 FastReasoning· LMArena data1,387.3
145Qwen3.5 122B-A10B· LMArena data1,386.9
146Claude Haiku 4.520251001· LMArena data1,385.9
147Grok 4.3· LMArena data1,384.6
148Hunyuan T120250711· LMArena data1,384.5
149MiMo V2 FlashNon-Thinking· LMArena data1,383.6
150GPT-5 Chat· LMArena data1,383.6

Showing 101–150 of 422 results.

Data credits: LMArena, CC BY 4.0 (licence)