Skip to content

Benchmark

Arena ELO - Exclude Ties

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

Models measured
366
Scores recorded
423
Sources
1
IQ weight
1.28

Score distribution

1,247.7
Mean
1,316.3
Median
217.2
Std dev
908.5
10th percentile
1,465.2
90th percentile
533.7
Min
533.71,565.6
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.28 for chat models, measured over 366 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
201GPT OSS 120B· LMArena data1,330.3
202Grok 3 Minibeta· LMArena data1,329.6
203Granite 4.2 30B· LMArena data1,329.1
204nova-experimental-chat10-09· LMArena data1,327.9
205MiniMax M2.5· LMArena data1,326.4
206Nova 2 Lite· LMArena data1,325.5
207Ling Flash 2.0· LMArena data1,323.3
208Gemma 3 27B IT· LMArena data1,321.3
209Qwen3 Coder· LMArena data1,318.3
210Gemini 2.0 Flash001· LMArena data1,317.5
211INTELLECT-3· LMArena data1,316.3
212o4 Mini2025-04-16· LMArena data1,314.6
213O1 Preview· LMArena data1,313.9
214Mercury 2· LMArena data1,312.4
215Claude Sonnet 4Thinking 32K, 20250514· LMArena data1,311.1
216Step3· LMArena data1,311
217GLM 4.7 Flash· LMArena data1,306.6
218Nemotron 3 Nano 30B A3BBF16· LMArena data1,303.2
219Trinity Large Thinking· LMArena data1,302.9
220Trinity Large Preview· LMArena data1,298.1
221MiniMax M1· LMArena data1,298.1
222GPT-4.1 mini2025-04-14· LMArena data1,296.3
223Qwen3 32B· LMArena data1,296.3
224MiniMax M2· LMArena data1,296.2
225Claude Sonnet 420250514· LMArena data1,295
226Llama 3.3 Nemotron Super 49B v1.5· LMArena data1,292.8
227Mistral Small 3.22506· LMArena data1,292.1
228o3 MiniHigh· LMArena data1,291.3
229Gemma 3 12B IT· LMArena data1,289
230DeepSeek V3· LMArena data1,288.6
231Step-1o Turbo202506· LMArena data1,288.4
232Ring Flash 2.0· LMArena data1,285.5
233GLM 4 Plus0111· LMArena data1,283.7
234Nemotron 3.5 Lightning 30B A3BNVFP4· LMArena data1,283.7
235GLM 4.5V· LMArena data1,282.3
236Command A03-2025· LMArena data1,281.1
237Gemini 2.0 Flash-Lite Preview02-05· LMArena data1,280.8
238Qwen Plus0125· LMArena data1,280
239QwQ 32B· LMArena data1,279.1
240Granite 4.2 8B· LMArena data1,272.3
241Step-2 16K Exp202412· LMArena data1,272
242GPT-5 nanoHigh· LMArena data1,268.3
243Gemini 1.5 Pro002· LMArena data1,265.9
244o3 Mini· LMArena data1,265.6
245Hunyuan TurboS20250226· LMArena data1,265.2
246Llama 3.1 Nemotron Ultra 253B v1· LMArena data1,263.4
247O1 Mini· LMArena data1,262.5
248Claude Sonnet 3.7Thinking 32K, 20250219· LMArena data1,260.9
249Qwen3 30B A3B· LMArena data1,259.7
250Llama 3.3 Nemotron Super 49B v1· LMArena data1,251.7

Showing 201–250 of 423 results.

Data credits: LMArena, CC BY 4.0 (licence)