Skip to content

Benchmark

Arena ELO - German

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

Models measured
273
Scores recorded
319
Sources
1
IQ weight
1.52

Score distribution

1,328.6
Mean
1,372.1
Median
138.3
Std dev
1,110.8
10th percentile
1,469.8
90th percentile
901.9
Min
901.91,585.2
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.52 for chat models, measured over 273 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
201Qwen3 30B A3B· LMArena data1,306.5
202Gemini 2.0 Flash-Lite Preview02-05· LMArena data1,305.5
203Nemotron 3.5 Lightning 30B A3BNVFP4· LMArena data1,304
204o3 MiniHigh· LMArena data1,303
205Claude Sonnet 3.7Thinking 32K, 20250219· LMArena data1,300.7
206o3 Mini· LMArena data1,298.9
207Claude 3.5 Sonnet20241022· LMArena data1,297.4
208Llama 4 Maverick· LMArena data1,290.7
209Olmo 3 32B Think· LMArena data1,290.2
210GPT-4.1 nano2025-04-14· LMArena data1,287.8
211Grok 22024-08-13· LMArena data1,286.8
212Gemini 1.5 Pro002· LMArena data1,286.3
213GPT 4o2024-05-13· LMArena data1,282.4
214Olmo 3.1 32B Instruct· LMArena data1,281.6
215Claude Sonnet 3.720250219· LMArena data1,281.1
216Gemma 3 4B IT· LMArena data1,280.7
217GLM 4 Plus· LMArena data1,278.4
218O1 Mini· LMArena data1,278
219Gemini Advanced0514· LMArena data1,274.8
220Grok 2 mini2024-08-13· LMArena data1,272.8
221Llama 4 Scout· LMArena data1,272.7
222GPT-4o mini2024-07-18· LMArena data1,272
223Claude 3.5 Sonnet20240620· LMArena data1,270.8
224Yi Lightning· LMArena data1,266.7
225Mistral Small 3.1 24B Instruct2503· LMArena data1,265.6
226Olmo 3.1 32B Think· LMArena data1,261.8
227Gemini 1.5 Flash002· LMArena data1,261.5
228Athene V2 Chat· LMArena data1,260.6
229GPT-4 Turbo2024-04-09· LMArena data1,259.4
230Claude 3 Opus20240229· LMArena data1,258.3
231DeepSeek V2.51210· LMArena data1,258.1
232GPT 4o2024-08-06· LMArena data1,257.3
233GPT OSS 20b· LMArena data1,254.9
234Mistral Large 22407· LMArena data1,253.9
235Qwen Max0919· LMArena data1,253.9
236Llama 3.1 405B InstructBF16· LMArena data1,252.3
237Athene 70B0725· LMArena data1,251.6
238Llama 3.3 70B Instruct· LMArena data1,250.8
239GPT-4 Turbo Preview1106· LMArena data1,250.5
240Llama 3.1 405B InstructFP8· LMArena data1,250.5
241Gemini 1.5 Pro001· LMArena data1,249.1
242Magistral Medium2506· LMArena data1,248.3
243GPT-4 Turbo Preview0125· LMArena data1,245.5
244Reka Core20240904· LMArena data1,243.2
245Nova Pro 1.0· LMArena data1,242.9
246Mistral Large2411· LMArena data1,241.8
247Claude Haiku 3.520241022· LMArena data1,236.5
248Qwen2.5 72B Instruct· LMArena data1,233.5
249Qwen2.5-Plus1127· LMArena data1,230.8
250Nova Lite 1.0· LMArena data1,228.6

Showing 201–250 of 319 results.

Data credits: LMArena, CC BY 4.0 (licence)