Skip to content

Benchmark

Arena ELO - French

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

Models measured
262
Scores recorded
303
Sources
1
IQ weight
1.87

Score distribution

1,360.8
Mean
1,408.4
Median
127.5
Std dev
1,158.9
10th percentile
1,487.2
90th percentile
945.6
Min
945.61,562.9
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.87 for chat models, measured over 262 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
201QwQ 32B· LMArena data1,338.2
202Mistral Small 3.22506· LMArena data1,336.5
203MiniMax M2· LMArena data1,335.4
204o3 MiniHigh· LMArena data1,334
205Athene V2 Chat· LMArena data1,333
206GLM 4.7 Flash· LMArena data1,332.3
207Qwen Max0919· LMArena data1,330.1
208Gemma 3n E4B IT· LMArena data1,329.9
209Olmo 3.1 32B Instruct· LMArena data1,328.7
210Mistral Large2411· LMArena data1,325.1
211Grok 22024-08-13· LMArena data1,317.7
212o3 Mini· LMArena data1,311.9
213GLM 4 Plus· LMArena data1,308.1
214Gemini Advanced0514· LMArena data1,306.6
215Yi Lightning· LMArena data1,305.3
216Claude 3.5 Sonnet20241022· LMArena data1,304.6
217GPT 4o2024-05-13· LMArena data1,304.1
218Claude Sonnet 3.720250219· LMArena data1,301.9
219Gemini 1.5 Pro002· LMArena data1,301.6
220GPT-4o mini2024-07-18· LMArena data1,296.8
221Claude 3.5 Sonnet20240620· LMArena data1,293.7
222O1 Mini· LMArena data1,293.2
223DeepSeek V2.5· LMArena data1,289.2
224Claude Sonnet 3.7Thinking 32K, 20250219· LMArena data1,289
225Athene 70B0725· LMArena data1,286.5
226Grok 2 mini2024-08-13· LMArena data1,286.2
227Olmo 3 32B Think· LMArena data1,285.8
228GPT-4 Turbo Preview1106· LMArena data1,282.8
229Llama 4 Scout· LMArena data1,282.4
230Llama 3.3 70B Instruct· LMArena data1,281.3
231GPT-4 Turbo Preview0125· LMArena data1,281.1
232Qwen2.5 72B Instruct· LMArena data1,280.3
233Llama 3.1 405B InstructBF16· LMArena data1,278.8
234GPT-4 Turbo2024-04-09· LMArena data1,275.6
235Claude 3 Opus20240229· LMArena data1,275.2
236Mistral Large 22407· LMArena data1,273
237Gemini 1.5 Pro001· LMArena data1,272.1
238Mistral Small 3.1 24B Instruct2503· LMArena data1,271.5
239Nova Pro 1.0· LMArena data1,271.2
240Llama 3.1 405B InstructFP8· LMArena data1,271
241Magistral Medium2506· LMArena data1,268.6
242Claude Haiku 3.520241022· LMArena data1,264.2
243GPT 4o2024-08-06· LMArena data1,262.8
244Olmo 3.1 32B Think· LMArena data1,261.7
245Llama 3.1 70B Instruct· LMArena data1,261.3
246Llama 4 Maverick· LMArena data1,259.8
247Gemini 1.5 Flash002· LMArena data1,258.3
248Aya Expanse 32B· LMArena data1,248
249Gemma 2 27B IT· LMArena data1,247.5
250Gemini 1.5 Flash001· LMArena data1,240.9

Showing 201–250 of 303 results.

Data credits: LMArena, CC BY 4.0 (licence)