Skip to content

Benchmark

Arena ELO - Creative Writing

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

Models measured
364
Scores recorded
421
Sources
1
IQ weight
1.14

Score distribution

1,287.4
Mean
1,318.1
Median
144.2
Std dev
1,079.9
10th percentile
1,442.3
90th percentile
793.9
Min
793.91,532.7
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.14 for chat models, measured over 364 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
251GPT OSS 120B· LMArena data1,274.5
252o3 Mini· LMArena data1,272.7
253Gemini 1.5 Pro001· LMArena data1,272.4
254Qwen3 30B A3B· LMArena data1,271.6
255Gemma 3 4B IT· LMArena data1,270.4
256Hunyuan TurboS20250226· LMArena data1,270.3
257Ring Flash 2.0· LMArena data1,269.8
258Nemotron 3 Nano 30B A3BBF16· LMArena data1,269.2
259Hunyuan Turbo0110· LMArena data1,268.9
260Llama 3.1 Nemotron 70B Instruct· LMArena data1,268.5
261GPT-4 Turbo2024-04-09· LMArena data1,268.4
262GPT-4o mini2024-07-18· LMArena data1,267.7
263Llama 4 Maverick· LMArena data1,266.2
264Olmo 3.1 32B Instruct· LMArena data1,264.7
265Hunyuan Large2025-02-10· LMArena data1,264.2
266GLM 4 Plus· LMArena data1,263.5
267Llama 3.1 405B InstructFP8· LMArena data1,261.9
268Qwen2.5-Plus1127· LMArena data1,261.8
269GPT-4.1 nano2025-04-14· LMArena data1,260
270Llama 3.1 405B InstructBF16· LMArena data1,259.6
271Nemotron 3.5 Lightning 30B A3BNVFP4· LMArena data1,259.5
272Granite 4.1 8B· LMArena data1,254.6
273Olmo 3 32B Think· LMArena data1,254.5
274Mistral Small 3.1 24B Instruct2503· LMArena data1,252.6
275Llama 3.3 70B Instruct· LMArena data1,250.4
276GPT-5 nanoHigh· LMArena data1,249.8
277Llama 4 Scout· LMArena data1,249
278Qwen Max0919· LMArena data1,248
279Hunyuan Large Vision· LMArena data1,247.7
280Magistral Medium2506· LMArena data1,244.7
281GPT-4 Turbo Preview1106· LMArena data1,244.2
282O1 Mini· LMArena data1,244.2
283Mistral Large 22407· LMArena data1,242.7
284Hunyuan Standard2025-02-10· LMArena data1,242.1
285Mistral Large2411· LMArena data1,241.7
286Grok 2 mini2024-08-13· LMArena data1,241.5
287Gemma 2 27B IT· LMArena data1,240.5
288Gemma 2 9B IT SimPO· LMArena data1,238.7
289Athene 70B0725· LMArena data1,238.2
290Claude 3.5 Sonnet20240620· LMArena data1,238
291DeepSeek V2.5· LMArena data1,236.7
292Reka Core20240904· LMArena data1,236.1
293Granite 4.2 8B· LMArena data1,235.2
294Claude 3 Opus20240229· LMArena data1,234.7
295Command R+08-2024· LMArena data1,234.6
296Athene V2 Chat· LMArena data1,234.2
297Claude Haiku 3.520241022· LMArena data1,232.5
298GPT-4 Turbo Preview0125· LMArena data1,232.3
299Llama 3.1 70B Instruct· LMArena data1,231.8
300Llama 3.1 Tulu 3 70B· LMArena data1,231.2

Showing 251–300 of 421 results.

Data credits: LMArena, CC BY 4.0 (licence)