Skip to content

Benchmark

Arena ELO - Spanish

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

Models measured
259
Scores recorded
305
Sources
1
IQ weight
1.90

Score distribution

1,355.9
Mean
1,399.3
Median
117.6
Std dev
1,167.6
10th percentile
1,470
90th percentile
1,007
Min
1,0071,522.3
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.90 for chat models, measured over 259 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
251Claude 3 Opus20240229· LMArena data1,246
252Gemini 1.5 Pro001· LMArena data1,242.8
253Gemini 1.5 Flash002· LMArena data1,242.3
254Gemini Advanced0514· LMArena data1,241.8
255Llama 3 70B Instruct· LMArena data1,240.8
256Mistral Large 22407· LMArena data1,236
257Phi 4· LMArena data1,234
258Gemma 2 27B IT· LMArena data1,228
259Nova Lite 1.0· LMArena data1,225.4
260Nova Micro 1.0· LMArena data1,224.8
261Gemini 1.5 Flash001· LMArena data1,224
262Gemini 1.5 Flash-8B001· LMArena data1,212
263Claude 3 Sonnet20240229· LMArena data1,204
264Gemma 2 9B IT· LMArena data1,200
265GPT 40314· LMArena data1,196.4
266Nemotron-4 340B Instruct· LMArena data1,194.9
267Mistral Large2402· LMArena data1,194.4
268Aya Expanse 32B· LMArena data1,192.7
269Command R+· LMArena data1,188.8
270Nova Pro 1.0· LMArena data1,181.9
271Llama 3 8B Instruct· LMArena data1,172.5
272Llama 3.1 8B Instruct· LMArena data1,169
273Qwen2 72B Instruct· LMArena data1,168.6
274GPT 40613· LMArena data1,167.6
275Claude 3 Haiku20240307· LMArena data1,165.3
276DeepSeek Coder V2 Instruct· LMArena data1,152.9
277Mixtral 8x22B Instruct v0.1· LMArena data1,150.4
278Reka Flash 21BOnline, 20240226· LMArena data1,148.7
279Mistral Medium· LMArena data1,143.1
280Llama 2 70B Chat· LMArena data1,142.2
281Qwen1.5-110B-Chat· LMArena data1,141.6
282Gemma 2 2B IT· LMArena data1,139.7
283Reka Flash 21B20240226· LMArena data1,131.7
284Yi 1.5 34B Chat· LMArena data1,120.4
285GPT-3.5 Turbo0125· LMArena data1,120.2
286Gemini ProDev API· LMArena data1,118.6
287Phi-3 Small 8K Instruct· LMArena data1,110.8
288Mixtral 8x7B Instruct v0.1· LMArena data1,110.6
289Qwen1.5-72B-Chat· LMArena data1,109.9
290Phi-3 Medium 4K Instruct· LMArena data1,094.7
291GPT-3.5 Turbo1106· LMArena data1,092.4
292Qwen1.5-32B-Chat· LMArena data1,088.7
293Snowflake Arctic Instruct· LMArena data1,087.8
294Llama 2 13B Chat· LMArena data1,086.2
295Qwen1.5-14B-Chat· LMArena data1,085
296Phi-3 Mini 4K Instruct· LMArena data1,084.5
297Vicuna 33B· LMArena data1,079.6
298Vicuna 13B· LMArena data1,076.3
299Zephyr 7B Beta· LMArena data1,073.2
300Yi 34B Chat· LMArena data1,069.8

Showing 251–300 of 305 results.

Data credits: LMArena, CC BY 4.0 (licence)