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
301GPT-4 Turbo Preview0125· LMArena data1,251.2
302Gemini Advanced0514· LMArena data1,249.3
303OLMo 2 32B Instruct0325· LMArena data1,243.6
304Claude 3 Opus20240229· LMArena data1,243.3
305Mistral Small 24B Instruct2501· LMArena data1,243.1
306Llama 3.1 Tulu 3 70B· LMArena data1,236.5
307Jamba 1.5 Large· LMArena data1,231.1
308Llama 3 70B Instruct· LMArena data1,230.4
309Phi 4· LMArena data1,229.9
310Hunyuan Standard 256K· LMArena data1,227.9
311Nova Lite 1.0· LMArena data1,225.7
312Gemini 1.5 Flash001· LMArena data1,224.6
313GLM 40520· LMArena data1,218.6
314DeepSeek Coder V2 Instruct· LMArena data1,217.3
315Reka Core20240904· LMArena data1,215.6
316Gemini 1.5 Flash-8B001· LMArena data1,214.5
317GPT 40314· LMArena data1,208.7
318Ministral 8B2410· LMArena data1,206.1
319Gemma 2 27B IT· LMArena data1,205.8
320Qwen2 72B Instruct· LMArena data1,205.8
321Llama 3.1 Nemotron 51B Instruct· LMArena data1,205.1
322Nova Micro 1.0· LMArena data1,203.1
323Claude 3 Sonnet20240229· LMArena data1,202.4
324Nemotron-4 340B Instruct· LMArena data1,201.6
325Gemma 2 9B IT SimPO· LMArena data1,201.1
326Aya Expanse 32B· LMArena data1,195.8
327Llama 3.1 8B Instruct· LMArena data1,195.6
328Command R+08-2024· LMArena data1,194.3
329InternLM2.5 20B Chat· LMArena data1,188.8
330GPT 40613· LMArena data1,188.1
331Reka Flash20240904· LMArena data1,186.1
332Qwen1.5-110B-Chat· LMArena data1,184.9
333Jamba 1.5 Mini· LMArena data1,181.8
334Claude 3 Haiku20240307· LMArena data1,181.4
335Yi 1.5 34B Chat· LMArena data1,179.6
336Mistral Large2402· LMArena data1,178.2
337Llama 3.1 Tulu 3 8B· LMArena data1,177.8
338Gemma 2 9B IT· LMArena data1,177.7
339Command R+· LMArena data1,171.3
340QwQ 32B Preview· LMArena data1,170.9
341Llama 3 8B Instruct· LMArena data1,163.7
342Granite 3.1 8B Instruct· LMArena data1,163.5
343Mixtral 8x22B Instruct v0.1· LMArena data1,163
344Qwen1.5-72B-Chat· LMArena data1,162.2
345Mistral Medium· LMArena data1,157.8
346Granite 3.1 2B Instruct· LMArena data1,156.4
347Aya Expanse 8B· LMArena data1,153.8
348Reka Flash 21BOnline, 20240226· LMArena data1,146.8
349Qwen1.5-32B-Chat· LMArena data1,146.6
350Reka Flash 21B20240226· LMArena data1,144.2

Showing 301–350 of 421 results.

Data credits: LMArena, CC BY 4.0 (licence)