Skip to content

Benchmark

Arena ELO - Industry Entertainment and Sports and Media

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

Models measured
364
Scores recorded
421
Sources
1
IQ weight
1.17

Score distribution

1,279.8
Mean
1,314.7
Median
149.5
Std dev
1,058.5
10th percentile
1,441.1
90th percentile
784.8
Min
784.81,526.7
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.17 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
351Yi 34B Chat· LMArena data1,106.7
352Mixtral 8x7B Instruct v0.1· LMArena data1,105.2
353Granite 3.1 2B Instruct· LMArena data1,104.7
354InternLM2.5 20B Chat· LMArena data1,103.1
355Vicuna 33B· LMArena data1,102.9
356Reka Flash 21B20240226· LMArena data1,100.6
357Gemini ProDev API· LMArena data1,099.7
358DBRX Instruct Preview· LMArena data1,098.8
359OpenChat 3.5· LMArena data1,096.4
360GPT-3.5 Turbo0125· LMArena data1,088.6
361Gemini Pro· LMArena data1,085
362Starling LM 7B Beta· LMArena data1,084.9
363Llama 3.2 3B· LMArena data1,083.4
364Starling LM 7B Alpha· LMArena data1,082.4
365Phi-3 Medium 4K Instruct· LMArena data1,081.2
366Qwen1.5-32B-Chat· LMArena data1,080.5
367WizardLM 13B· LMArena data1,074.9
368OpenChat 3.50106· LMArena data1,074.3
369OpenHermes 2.5 Mistral 7B· LMArena data1,074
370Guanaco 33B· LMArena data1,072.8
371Llama 2 70B SteerLM Chat· LMArena data1,069.2
372Snowflake Arctic Instruct· LMArena data1,068.9
373Qwen1.5-14B-Chat· LMArena data1,068.4
374Nous Hermes 2 Mixtral 8x7B DPO· LMArena data1,066.8
375Llama 2 70B Chat· LMArena data1,064.5
376Zephyr 7B Beta· LMArena data1,062.7
377DeepSeek LLM 67B Chat· LMArena data1,061.7
378Granite 3.0 8B Instruct· LMArena data1,061.1
379SOLAR 10.7B Instruct v1.0· LMArena data1,058.5
380Falcon 180B Chat· LMArena data1,058.3
381Mistral 7B Instruct v0.2· LMArena data1,057.9
382Phi-3 Small 8K Instruct· LMArena data1,055.5
383MPT 30B Chat· LMArena data1,055.2
384Gemma 1.1 7B IT· LMArena data1,049.7
385Dolphin 2.2.1 Mistral 7B· LMArena data1,049.6
386Zephyr 7B Alpha· LMArena data1,048.3
387Llama 2 13B Chat· LMArena data1,047.5
388Granite 3.0 2B Instruct· LMArena data1,037.7
389GPT-3.5 Turbo1106· LMArena data1,032.6
390Phi-3 Mini 4K InstructJune 2024· LMArena data1,031.4
391Vicuna 13B· LMArena data1,030.9
392Qwen1.5-7B-Chat· LMArena data1,029.9
393Llama 3.2 1B Instruct· LMArena data1,025.1
394Llama 2 7B Chat· LMArena data1,019.8
395Phi-3 Mini 4K Instruct· LMArena data1,018.9
396Vicuna 7B· LMArena data1,016.6
397Code Llama 34B Instruct· LMArena data1,014.7
398StripedHyena Nous 7B· LMArena data1,011.4
399Phi-3 Mini 128K Instruct· LMArena data1,009.1
400SmolLM2 1.7B Instruct· LMArena data1,007.7

Showing 351–400 of 421 results.

Data credits: LMArena, CC BY 4.0 (licence)