Skip to content

Benchmark

Arena ELO - Russian

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

Models measured
332
Scores recorded
387
Sources
1
IQ weight
1.42

Score distribution

1,322.5
Mean
1,355.7
Median
134.6
Std dev
1,109.2
10th percentile
1,466.9
90th percentile
936.6
Min
936.61,560.5
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.42 for chat models, measured over 332 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
351Qwen1.5-72B-Chat· LMArena data1,104.1
352Snowflake Arctic Instruct· LMArena data1,103.4
353Starling LM 7B Beta· LMArena data1,102.5
354Starling LM 7B Alpha· LMArena data1,095.5
355Granite 3.1 8B Instruct· LMArena data1,091.8
356Yi 1.5 34B Chat· LMArena data1,090.9
357Mixtral 8x7B Instruct v0.1· LMArena data1,090.3
358OpenChat 3.50106· LMArena data1,084.7
359Llama 2 70B Chat· LMArena data1,082.8
360Vicuna 33B· LMArena data1,082.6
361DBRX Instruct Preview· LMArena data1,077.9
362GPT-3.5 Turbo1106· LMArena data1,077.3
363Qwen1.5-32B-Chat· LMArena data1,072.8
364Granite 3.1 2B Instruct· LMArena data1,063.3
365Zephyr 7B Beta· LMArena data1,062
366Granite 3.0 8B Instruct· LMArena data1,060.3
367Llama 2 13B Chat· LMArena data1,054.5
368Yi 34B Chat· LMArena data1,049.4
369Gemma 1.1 7B IT· LMArena data1,045.8
370Qwen1.5-14B-Chat· LMArena data1,045.8
371Granite 3.0 2B Instruct· LMArena data1,045.2
372Vicuna 13B· LMArena data1,042.1
373Phi-3 Mini 4K Instruct· LMArena data1,022.2
374Phi-3 Mini 4K InstructJune 2024· LMArena data1,019.1
375Mistral 7B Instruct v0.2· LMArena data1,018.2
376Qwen1.5-7B-Chat· LMArena data1,005.7
377Phi-3 Mini 128K Instruct· LMArena data1,003.9
378Mistral 7B Instruct· LMArena data1,001.3
379Llama 2 7B Chat· LMArena data995
380Gemma 7B IT· LMArena data993.2
381Gemma 1.1 2B IT· LMArena data990.3
382SmolLM2 1.7B Instruct· LMArena data981.2
383Qwen1.5-4B-Chat· LMArena data952
384Llama 3.2 3B· LMArena data948.9
385OLMo 7B Instruct· LMArena data946.4
386Llama 3.2 1B Instruct· LMArena data941.1
387Gemma 2B IT· LMArena data936.6

Showing 351–387 of 387 results.

Data credits: LMArena, CC BY 4.0 (licence)