Skip to content

Benchmark

Arena ELO - Longer Query

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

Models measured
345
Scores recorded
401
Sources
1
IQ weight
1.34

Score distribution

1,323.7
Mean
1,357.1
Median
137.1
Std dev
1,102.6
10th percentile
1,475.3
90th percentile
951.6
Min
951.61,549
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 345 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
101Gemma 4 26B A4B· LMArena data1,427.9
102GPT-6.2 SolMax· LMArena data1,427.4
103GPT-5.2 Chat20260210· LMArena data1,426.9
104Grok 4.20 Beta 1· LMArena data1,426.8
105Beam 1.0 501B A23B· LMArena data1,426.8
106Qwen3 235B A22B Instruct 25072507· LMArena data1,426.2
107Claude Haiku 4.520251001· LMArena data1,426.2
108DeepSeek V3.2Thinking· LMArena data1,426.2
109LongCat-Flash-Chat Exp2602· LMArena data1,424.6
110DeepSeek-V4-Flash PreviewHigh· LMArena data1,424.5
111GLM 4.6· LMArena data1,422.5
112Claude Opus 4Thinking 16K, 20250514· LMArena data1,422
113GPT-5.5 Instant· LMArena data1,422
114DeepSeek V3.1Thinking· LMArena data1,421.5
115DeepSeek V3.1 TerminusThinking· LMArena data1,421.4
116ERNIE 5.00110· LMArena data1,421
117Qwen3 VL 235B· LMArena data1,419.5
118Gemini 2.5 Flash· LMArena data1,418.9
119MiniMax M2.7· LMArena data1,418.2
120DeepSeek V3.2 Exp· LMArena data1,417.9
121Grok 4.7xHigh· LMArena data1,416.2
122Grok 4.1· LMArena data1,415.9
123Kimi K2 Thinking· LMArena data1,415.5
124MiniMax-M2.1 Preview· LMArena data1,415.3
125GPT-6 SolMax· LMArena data1,415.2
126Mistral Medium 3.5· LMArena data1,415
127Mistral Large 3· LMArena data1,414.8
128Grok 4 FastNon-Reasoning· LMArena data1,413.8
129Qwen3.5 27B· LMArena data1,413.5
130ChatGPT-4o Latest20250326· LMArena data1,412.7
131GPT 5.2· LMArena data1,412.4
132GLM 4.5· LMArena data1,412.3
133Grok 4 FastReasoning· LMArena data1,412
134DeepSeek V3.2 ExpThinking· LMArena data1,410.8
135Qwen3.5 122B-A10B· LMArena data1,410
136Qwen3 Max2025-09-23· LMArena data1,409.9
137MiMo V2 FlashNon-Thinking· LMArena data1,409.6
138Mistral Large 4· LMArena data1,409.2
139GPT 5.2High· LMArena data1,408.5
140ERNIE-5.0-Preview1203· LMArena data1,408.3
141Grok 40709· LMArena data1,408.1
142GPT-5.4 miniHigh· LMArena data1,407.5
143Grok 4.1Thinking· LMArena data1,406.7
144Mistral Medium 3.12508· LMArena data1,405.9
145ERNIE-5.0-Preview1022· LMArena data1,405.9
146Hunyuan Vision 1.5 Thinking· LMArena data1,405.8
147Gemini 2.5 Flash Preview09-2025· LMArena data1,405.6
148GPT-4.5 Preview2025-02-27· LMArena data1,405.6
149GPT-6 LunaMax· LMArena data1,404.2
150Inkling Small· LMArena data1,403.1

Showing 101–150 of 401 results.

Data credits: LMArena, CC BY 4.0 (licence)