Skip to content

Benchmark

Arena ELO - Chinese

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

Models measured
340
Scores recorded
396
Sources
1
IQ weight
1.23

Score distribution

1,334.6
Mean
1,384.4
Median
172.2
Std dev
1,067.9
10th percentile
1,512.4
90th percentile
719
Min
7191,628.9
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.23 for chat models, measured over 340 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
301Gemma 2 27B IT· LMArena data1,221
302Reka Flash20240904· LMArena data1,220.8
303QwQ 32B Preview· LMArena data1,217.9
304Llama 3.3 70B Instruct· LMArena data1,217.1
305Llama 3.1 70B Instruct· LMArena data1,215.3
306Nemotron-4 340B Instruct· LMArena data1,213.9
307Yi 1.5 34B Chat· LMArena data1,213.1
308Phi 4· LMArena data1,211.5
309Aya Expanse 32B· LMArena data1,210.6
310Nova Micro 1.0· LMArena data1,209.4
311Qwen1.5-110B-Chat· LMArena data1,205.9
312Mistral Small 24B Instruct2501· LMArena data1,203.7
313Jamba 1.5 Large· LMArena data1,203.1
314DeepSeek Coder V2 Instruct· LMArena data1,200.9
315InternLM2.5 20B Chat· LMArena data1,199.5
316Ministral 8B2410· LMArena data1,193.5
317OLMo 2 32B Instruct0325· LMArena data1,191.5
318Command R+· LMArena data1,191.1
319Claude 3 Sonnet20240229· LMArena data1,188.7
320Qwen1.5-72B-Chat· LMArena data1,186.3
321Gemma 2 9B IT· LMArena data1,185
322GPT 40314· LMArena data1,184
323Aya Expanse 8B· LMArena data1,181
324Llama 3.1 Nemotron 51B Instruct· LMArena data1,179.7
325Qwen1.5-32B-Chat· LMArena data1,176.6
326Llama 3.1 Tulu 3 8B· LMArena data1,175.7
327Yi 34B Chat· LMArena data1,175.7
328Claude 3 Haiku20240307· LMArena data1,154.8
329Llama 3.1 8B Instruct· LMArena data1,150.9
330Qwen1.5-14B-Chat· LMArena data1,147.1
331Granite 3.1 8B Instruct· LMArena data1,145
332Reka Flash 21BOnline, 20240226· LMArena data1,143.3
333Qwen1.5-7B-Chat· LMArena data1,140.6
334Granite 3.1 2B Instruct· LMArena data1,138.6
335GPT 40613· LMArena data1,136.2
336Jamba 1.5 Mini· LMArena data1,135.7
337Reka Flash 21B20240226· LMArena data1,135.4
338Gemma 2 2B IT· LMArena data1,132.2
339DeepSeek LLM 67B Chat· LMArena data1,131.9
340Gemini ProDev API· LMArena data1,124.1
341Mistral Large2402· LMArena data1,120.4
342Mixtral 8x22B Instruct v0.1· LMArena data1,116.4
343Llama 3 70B Instruct· LMArena data1,114.3
344Starling LM 7B Beta· LMArena data1,114
345Mistral Medium· LMArena data1,109.2
346Phi-3 Medium 4K Instruct· LMArena data1,108.2
347Gemini Pro· LMArena data1,096.5
348OpenChat 3.50106· LMArena data1,083.8
349Zephyr ORPO 141B A35B v0.1· LMArena data1,082.1
350ChatGLM 6B· LMArena data1,077.4

Showing 301–350 of 396 results.

Data credits: LMArena, CC BY 4.0 (licence)