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
351Phi-3 Medium 4K Instruct· LMArena data1,142.4
352Llama 3.2 3B· LMArena data1,140.8
353DBRX Instruct Preview· LMArena data1,137.6
354Mixtral 8x7B Instruct v0.1· LMArena data1,134.9
355Zephyr ORPO 141B A35B v0.1· LMArena data1,133
356Starling LM 7B Beta· LMArena data1,130.9
357Gemma 2 2B IT· LMArena data1,127.2
358Phi-3 Small 8K Instruct· LMArena data1,125.2
359Tulu 2 DPO 70B· LMArena data1,124.2
360Qwen1.5-14B-Chat· LMArena data1,123.6
361GPT-3.5 Turbo0125· LMArena data1,121.8
362Granite 3.0 8B Instruct· LMArena data1,121.5
363Yi 34B Chat· LMArena data1,120.1
364Gemini ProDev API· LMArena data1,116.8
365Phi-3 Mini 4K InstructJune 2024· LMArena data1,114.5
366GPT-3.5 Turbo1106· LMArena data1,114.2
367Gemini Pro· LMArena data1,110.6
368Phi-3 Mini 4K Instruct· LMArena data1,097.7
369WizardLM 70B· LMArena data1,096.1
370OpenChat 3.50106· LMArena data1,094.9
371Starling LM 7B Alpha· LMArena data1,094.7
372Llama 2 70B Chat· LMArena data1,091.3
373Gemma 1.1 7B IT· LMArena data1,090.7
374Mistral 7B Instruct v0.2· LMArena data1,089.5
375Snowflake Arctic Instruct· LMArena data1,088.1
376Granite 3.0 2B Instruct· LMArena data1,086.9
377Llama 3.2 1B Instruct· LMArena data1,086.7
378OpenHermes 2.5 Mistral 7B· LMArena data1,086.6
379DeepSeek LLM 67B Chat· LMArena data1,085.3
380MPT 30B Chat· LMArena data1,082.4
381Vicuna 33B· LMArena data1,079.6
382Code Llama 70B Instruct· LMArena data1,077.5
383SOLAR 10.7B Instruct v1.0· LMArena data1,075.2
384Qwen1.5-7B-Chat· LMArena data1,075.2
385Llama 2 13B Chat· LMArena data1,072.7
386OpenChat 3.5· LMArena data1,071.1
387SmolLM2 1.7B Instruct· LMArena data1,068.4
388Dolphin 2.2.1 Mistral 7B· LMArena data1,068.1
389Nous Hermes 2 Mixtral 8x7B DPO· LMArena data1,066.4
390Gemma 7B IT· LMArena data1,060.6
391Llama 2 70B SteerLM Chat· LMArena data1,056.7
392Phi-3 Mini 128K Instruct· LMArena data1,051.2
393Zephyr 7B Beta· LMArena data1,047.7
394Code Llama 34B Instruct· LMArena data1,046.9
395Zephyr 7B Alpha· LMArena data1,042
396Qwen-14B-Chat· LMArena data1,041
397Vicuna 13B· LMArena data1,035.6
398PaLM 2· LMArena data1,035.3
399WizardLM 13B· LMArena data1,033.8
400Llama 2 7B Chat· LMArena data1,031.6

Showing 351–400 of 421 results.

Data credits: LMArena, CC BY 4.0 (licence)