Skip to content

Benchmark

Arena ELO - Hard Prompts English

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

Models measured
364
Scores recorded
421
Sources
1
IQ weight
1.31

Score distribution

1,319.1
Mean
1,374
Median
157.7
Std dev
1,085.3
10th percentile
1,479.2
90th percentile
734.7
Min
734.71,547.3
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.31 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
201Hunyuan TurboS20250416· LMArena data1,376.8
202MiniMax M2· LMArena data1,376.6
203INTELLECT-3· LMArena data1,375.9
204Gemini 2.5 Flash-Lite PreviewThinking, 06-17· LMArena data1,375.7
205O1 Preview· LMArena data1,375.6
206Grok 3 Minibeta· LMArena data1,375.6
207o3 MiniHigh· LMArena data1,375.4
208Nemotron 3 Nano 30B A3BBF16· LMArena data1,375
209DeepSeek R1· LMArena data1,374.6
210Mercury 2· LMArena data1,374
211Mistral Medium 32505· LMArena data1,373.8
212Nemotron 3.5 Lightning 30B A3BNVFP4· LMArena data1,372.3
213o12024-12-17· LMArena data1,372.1
214Qwen 3 235B A22B· LMArena data1,371.8
215Kimi K2 Preview0711· LMArena data1,369.5
216GLM 4.7 Flash· LMArena data1,368.9
217GPT OSS 120B· LMArena data1,367.9
218Trinity Large Preview· LMArena data1,367.2
219o4 Mini2025-04-16· LMArena data1,365.5
220nova-experimental-chat10-09· LMArena data1,364.8
221GPT-4.1 mini2025-04-14· LMArena data1,362.1
222Qwen2.5-Max· LMArena data1,361.7
223Claude Sonnet 420250514· LMArena data1,358.6
224Mistral Small 3.22506· LMArena data1,357.1
225GLM 4.5V· LMArena data1,356.8
226Trinity Large Thinking· LMArena data1,355.8
227Gemini 2.0 Flash001· LMArena data1,355.4
228Step-1o Turbo202506· LMArena data1,354.3
229Gemma 3 27B IT· LMArena data1,353.4
230MiniMax M1· LMArena data1,353.1
231Olmo 3.1 32B Instruct· LMArena data1,352.3
232O1 Mini· LMArena data1,351.3
233Qwen3 32B· LMArena data1,349.5
234Llama 3.3 Nemotron Super 49B v1.5· LMArena data1,344
235o3 Mini· LMArena data1,342.9
236Granite 4.2 8B· LMArena data1,341.6
237QwQ 32B· LMArena data1,340.8
238Claude Sonnet 3.7Thinking 32K, 20250219· LMArena data1,340
239GPT-5 nanoHigh· LMArena data1,338
240Llama 3.3 Nemotron Super 49B v1· LMArena data1,337.7
241Hunyuan TurboS20250226· LMArena data1,337.2
242Llama 3.1 Nemotron Ultra 253B v1· LMArena data1,336.2
243Olmo 3 32B Think· LMArena data1,335.8
244Command A03-2025· LMArena data1,334.8
245Qwen Plus0125· LMArena data1,332.9
246Granite 4.2 3B· LMArena data1,332.7
247Gemini 2.0 Flash-Lite Preview02-05· LMArena data1,329.8
248Qwen3 30B A3B· LMArena data1,328.8
249Olmo 3.1 32B Think· LMArena data1,327.7
250Hunyuan Turbo0110· LMArena data1,326.8

Showing 201–250 of 421 results.

Data credits: LMArena, CC BY 4.0 (licence)