Skip to content

Benchmark

Arena ELO - Expert

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

Models measured
318
Scores recorded
371
Sources
1
IQ weight
1.30

Score distribution

1,333.9
Mean
1,371.6
Median
146.5
Std dev
1,113
10th percentile
1,491.8
90th percentile
953.7
Min
953.71,581.3
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.30 for chat models, measured over 318 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
201Nemotron 3 Nano 30B A3BBF16· LMArena data1,356.3
202GPT OSS 120B· LMArena data1,355.1
203Ling Flash 2.0· LMArena data1,353.8
204Trinity Large Preview· LMArena data1,352.5
205GLM 4.5V· LMArena data1,351.4
206DeepSeek V30324· LMArena data1,350.2
207Qwen 3 235B A22B· LMArena data1,349.2
208Kimi K2 Preview0711· LMArena data1,345.8
209Ring Flash 2.0· LMArena data1,343
210o4 Mini2025-04-16· LMArena data1,342.5
211Mistral Medium 32505· LMArena data1,341.7
212MiniMax M2· LMArena data1,340
213Gemini 2.0 Flash001· LMArena data1,339.1
214Hunyuan TurboS20250416· LMArena data1,338.5
215GPT-4.1 mini2025-04-14· LMArena data1,338.3
216Qwen3 Coder· LMArena data1,337.9
217O1 Preview· LMArena data1,337.5
218DeepSeek R1· LMArena data1,337.4
219Qwen2.5-Max· LMArena data1,336.5
220Claude Sonnet 420250514· LMArena data1,335.6
221Step3· LMArena data1,331.1
222o3 Mini· LMArena data1,329.8
223Llama 3.3 Nemotron Super 49B v1.5· LMArena data1,329.7
224Qwen Plus0125· LMArena data1,327.6
225QwQ 32B· LMArena data1,323.3
226INTELLECT-3· LMArena data1,322.7
227GPT-5 nanoHigh· LMArena data1,320.9
228Claude Sonnet 3.7Thinking 32K, 20250219· LMArena data1,320.5
229MiniMax M1· LMArena data1,317.2
230O1 Mini· LMArena data1,315.6
231Qwen3 30B A3B· LMArena data1,313.9
232Granite 4.1 8B· LMArena data1,311.2
233Granite 4.2 3B· LMArena data1,308.6
234Step-1o Turbo202506· LMArena data1,307.4
235Olmo 3.1 32B Instruct· LMArena data1,306.4
236DeepSeek V3· LMArena data1,305.4
237Gemini 2.0 Flash-Lite Preview02-05· LMArena data1,304.4
238Gemma 3 27B IT· LMArena data1,303.5
239Claude Sonnet 3.720250219· LMArena data1,300.6
240Command A03-2025· LMArena data1,295
241Olmo 3.1 32B Think· LMArena data1,294.3
242Mistral Small 3.22506· LMArena data1,291.3
243Qwen2.5-Plus1127· LMArena data1,289
244Yi Lightning· LMArena data1,285.3
245GLM 4 Plus0111· LMArena data1,279.1
246Step-2 16K Exp202412· LMArena data1,278.4
247Gemini 1.5 Pro002· LMArena data1,278.3
248Hunyuan Large2025-02-10· LMArena data1,275.4
249Olmo 3 32B Think· LMArena data1,273.1
250GPT-4.1 nano2025-04-14· LMArena data1,271.3

Showing 201–250 of 371 results.

Data credits: LMArena, CC BY 4.0 (licence)