Skip to content

Benchmark

Arena ELO - Industry Legal and Government

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

Models measured
338
Scores recorded
394
Sources
1
IQ weight
1.34

Score distribution

1,329.7
Mean
1,368.1
Median
139.4
Std dev
1,104.4
10th percentile
1,475.2
90th percentile
895.4
Min
895.41,543.1
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.34 for chat models, measured over 338 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
101GPT-5.2 Chat20260210· LMArena data1,437.2
102Mistral Medium 3.5· LMArena data1,436.8
103Qwen3.6 Plus· LMArena data1,436.5
104Qwen3 Max Preview· LMArena data1,434.7
105nova-experimental-chat12-10· LMArena data1,434.4
106ERNIE 5.00110· LMArena data1,434.1
107Grok 4.20 Beta 1· LMArena data1,433.8
108GPT-4.5 Preview2025-02-27· LMArena data1,433.5
109Gemini 3.5 Flash Lite· LMArena data1,430.8
110Grok 4.6High· LMArena data1,430.5
111GPT-5 Chat· LMArena data1,430.3
112Gemini 2.5 Flash· LMArena data1,430.3
113Qwen3 Next 80B A3B Instruct· LMArena data1,430.3
114Hy3 (Deep Infra)· LMArena data1,429.8
115MiMo V2 Omni· LMArena data1,429.1
116Qwen3 235B A22B Instruct 25072507· LMArena data1,428.5
117OpenAI o32025-04-16· LMArena data1,428.4
118Grok 40709· LMArena data1,427.9
119GPT 5.2High· LMArena data1,427.7
120Gemini 2.5 Flash Preview09-2025· LMArena data1,427.6
121Beam 1.0 501B A23B· LMArena data1,427.6
122Mistral Large 4· LMArena data1,427.5
123Grok 4 FastReasoning· LMArena data1,427
124MiniMax M3· LMArena data1,427
125Kimi K2.5Instant· LMArena data1,426.2
126DeepSeek V3.2· LMArena data1,425.1
127nova-experimental-chat11-10· LMArena data1,424.3
128MiMo V2.5· LMArena data1,423.8
129GPT 5High· LMArena data1,423.4
130Qwen3.5 122B-A10B· LMArena data1,421.8
131Claude Opus 4.1Thinking 16K, 20250805· LMArena data1,420.4
132GPT 5.2· LMArena data1,420.3
133DeepSeek V3.2 ExpThinking· LMArena data1,419.3
134DeepSeek R10528· LMArena data1,419
135DeepSeek V3.2Thinking· LMArena data1,418.6
136GPT-5.5 Instant· LMArena data1,418.5
137Grok 4.7xHigh· LMArena data1,418.5
138Gemini 3.1 Flash Lite· LMArena data1,418.4
139LongCat Flash Chat· LMArena data1,417.5
140Claude Opus 4.120250805· LMArena data1,416.7
141DeepSeek V3.1Thinking· LMArena data1,416.7
142Hunyuan T120250711· LMArena data1,416.7
143Kimi K2 Thinking· LMArena data1,416.3
144LongCat-Flash-Chat Exp2602· LMArena data1,416.2
145DeepSeek V3.1· LMArena data1,416
146GPT-6.2 AstraMax· LMArena data1,415.2
147Qwen 3 235B A22B2507, thinking· LMArena data1,415
148GPT-5.4 miniHigh· LMArena data1,414.4
149Inkling Small· LMArena data1,413.5
150MiniMax-M2.1 Preview· LMArena data1,409.9

Showing 101–150 of 394 results.

Data credits: LMArena, CC BY 4.0 (licence)