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
351Qwen-14B-Chat· LMArena data1,077
352Llama 3 8B Instruct· LMArena data1,076.1
353GPT-3.5 Turbo0125· LMArena data1,075.4
354Granite 3.0 2B Instruct· LMArena data1,070.1
355OpenChat 3.5· LMArena data1,070.1
356DBRX Instruct Preview· LMArena data1,067.9
357Snowflake Arctic Instruct· LMArena data1,067.2
358ChatGLM3 6B· LMArena data1,064.2
359Granite 3.0 8B Instruct· LMArena data1,063.2
360Phi-3 Small 8K Instruct· LMArena data1,060.8
361Gemma 1.1 7B IT· LMArena data1,060.7
362Mixtral 8x7B Instruct v0.1· LMArena data1,055.5
363WizardLM 70B· LMArena data1,052.5
364Starling LM 7B Alpha· LMArena data1,045.3
365Vicuna 13B· LMArena data1,039.6
366Vicuna 33B· LMArena data1,035.9
367Gemma 7B IT· LMArena data1,035.5
368Phi-3 Mini 4K InstructJune 2024· LMArena data1,033.5
369Qwen1.5-4B-Chat· LMArena data1,024
370WizardLM 13B· LMArena data1,022.8
371Phi-3 Mini 4K Instruct· LMArena data1,021.1
372Llama 3.2 3B· LMArena data1,017.3
373Phi-3 Mini 128K Instruct· LMArena data1,016.4
374OpenHermes 2.5 Mistral 7B· LMArena data1,014.4
375OLMo 7B Instruct· LMArena data1,013.7
376Tulu 2 DPO 70B· LMArena data1,013.6
377GPT-3.5 Turbo1106· LMArena data1,012.2
378Gemma 1.1 2B IT· LMArena data1,011.6
379Mistral 7B Instruct v0.2· LMArena data1,009.4
380Llama 2 13B Chat· LMArena data1,001.1
381Llama 2 70B Chat· LMArena data994.7
382Gemma 2B IT· LMArena data986
383Vicuna 7B· LMArena data979.1
384Code Llama 34B Instruct· LMArena data976.5
385Llama 2 7B Chat· LMArena data973
386Zephyr 7B Beta· LMArena data960.1
387Llama 3.2 1B Instruct· LMArena data958.7
388MPT 7B Chat· LMArena data954.8
389Mistral 7B Instruct· LMArena data931.6
390RWKV-4 Raven 14B· LMArena data910.7
391PaLM 2· LMArena data887.3
392Koala 13B· LMArena data881
393Dolly 2.0 12B· LMArena data836.5
394OpenAssistant Pythia 12B· LMArena data803
395Alpaca 13B· LMArena data782
396FastChat-T5 3B· LMArena data719

Showing 351–396 of 396 results.

Data credits: LMArena, CC BY 4.0 (licence)