Skip to content

Benchmark

Arena ELO - Multi Turn

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

Models measured
364
Scores recorded
421
Sources
1
IQ weight
1.21

Score distribution

1,304.7
Mean
1,350
Median
160.8
Std dev
1,057.9
10th percentile
1,467.2
90th percentile
740.3
Min
740.31,569.6
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.21 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
351Gemma 2 2B IT· LMArena data1,117.9
352GPT-3.5 Turbo0125· LMArena data1,116.8
353Mixtral 8x7B Instruct v0.1· LMArena data1,114.6
354Yi 34B Chat· LMArena data1,112.7
355Starling LM 7B Beta· LMArena data1,111.2
356DBRX Instruct Preview· LMArena data1,111
357Qwen1.5-14B-Chat· LMArena data1,109.7
358WizardLM 70B· LMArena data1,108.3
359Granite 3.1 8B Instruct· LMArena data1,108
360Llama 3.2 3B· LMArena data1,105.1
361Zephyr ORPO 141B A35B v0.1· LMArena data1,099.5
362Granite 3.1 2B Instruct· LMArena data1,099.4
363OpenChat 3.50106· LMArena data1,090.8
364Llama 2 70B Chat· LMArena data1,088
365Phi-3 Medium 4K Instruct· LMArena data1,087.9
366Tulu 2 DPO 70B· LMArena data1,085.4
367OpenChat 3.5· LMArena data1,085.3
368Starling LM 7B Alpha· LMArena data1,085
369DeepSeek LLM 67B Chat· LMArena data1,081.9
370Nous Hermes 2 Mixtral 8x7B DPO· LMArena data1,081.1
371Vicuna 33B· LMArena data1,080.4
372GPT-3.5 Turbo1106· LMArena data1,077.8
373Snowflake Arctic Instruct· LMArena data1,072.2
374Phi-3 Small 8K Instruct· LMArena data1,068.4
375OpenHermes 2.5 Mistral 7B· LMArena data1,064.4
376Granite 3.0 8B Instruct· LMArena data1,062.8
377Mistral 7B Instruct v0.2· LMArena data1,062.4
378Qwen1.5-7B-Chat· LMArena data1,061.5
379Llama 2 70B SteerLM Chat· LMArena data1,057.9
380MPT 30B Chat· LMArena data1,054.5
381Granite 3.0 2B Instruct· LMArena data1,052.9
382Llama 2 13B Chat· LMArena data1,050.1
383Phi-3 Mini 4K InstructJune 2024· LMArena data1,049.1
384WizardLM 13B· LMArena data1,047
385SOLAR 10.7B Instruct v1.0· LMArena data1,045.1
386Gemma 1.1 7B IT· LMArena data1,039.8
387Zephyr 7B Beta· LMArena data1,037.9
388Vicuna 13B· LMArena data1,030.8
389Llama 3.2 1B Instruct· LMArena data1,029.8
390Llama 2 7B Chat· LMArena data1,029.1
391Qwen-14B-Chat· LMArena data1,022.1
392Dolphin 2.2.1 Mistral 7B· LMArena data1,020.6
393Phi-3 Mini 4K Instruct· LMArena data1,018.2
394Zephyr 7B Alpha· LMArena data1,017.6
395Code Llama 34B Instruct· LMArena data1,015.3
396Falcon 180B Chat· LMArena data1,012.9
397Mistral 7B Instruct· LMArena data1,008.1
398StripedHyena Nous 7B· LMArena data1,007.6
399OLMo 7B Instruct· LMArena data1,007
400Guanaco 33B· LMArena data1,005.9

Showing 351–400 of 421 results.

Data credits: LMArena, CC BY 4.0 (licence)