Skip to content

Benchmark

Arena ELO - Industry Legal and Government

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

Models measured
338
Scores recorded
394
Sources
1
IQ weight
1.34

Score distribution

1,329.7
Mean
1,368.1
Median
139.4
Std dev
1,104.4
10th percentile
1,475.2
90th percentile
895.4
Min
895.41,543.1
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.34 for chat models, measured over 338 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
251GPT-4o mini2024-07-18· LMArena data1,305.2
252o3 Mini· LMArena data1,305.1
253Gemini Advanced0514· LMArena data1,304.2
254Llama 3.1 405B InstructFP8· LMArena data1,303.3
255Llama 3.1 Nemotron 70B Instruct· LMArena data1,302.9
256Gemini 1.5 Flash002· LMArena data1,301.9
257Llama 3.1 405B InstructBF16· LMArena data1,301.8
258Claude Sonnet 3.7Thinking 32K, 20250219· LMArena data1,301.5
259Claude Sonnet 3.720250219· LMArena data1,300.1
260GPT-4.1 nano2025-04-14· LMArena data1,300
261Athene V2 Chat· LMArena data1,299.1
262GPT-5 nanoHigh· LMArena data1,297.8
263Athene 70B0725· LMArena data1,293.1
264Grok 2 mini2024-08-13· LMArena data1,292.3
265Hunyuan Standard2025-02-10· LMArena data1,289.7
266Qwen2.5 72B Instruct· LMArena data1,287.8
267Olmo 3 32B Think· LMArena data1,287.6
268Olmo 3.1 32B Think· LMArena data1,287.2
269Llama 4 Maverick· LMArena data1,285.8
270Llama 4 Scout· LMArena data1,285
271Llama 3.3 70B Instruct· LMArena data1,284.2
272Llama 3.1 70B Instruct· LMArena data1,283.4
273Qwen Max0919· LMArena data1,282.9
274DeepSeek V2.5· LMArena data1,282.1
275Nova Pro 1.0· LMArena data1,281.4
276Hunyuan Large2025-02-10· LMArena data1,280.6
277GPT OSS 20b· LMArena data1,280
278DeepSeek V2.51210· LMArena data1,278.6
279Claude 3.5 Sonnet20241022· LMArena data1,276.9
280Mistral Large 22407· LMArena data1,276
281GPT 4o2024-08-06· LMArena data1,275.2
282Mistral Large2411· LMArena data1,274.8
283Reka Core20240904· LMArena data1,273.4
284Mistral Small 3.1 24B Instruct2503· LMArena data1,270.3
285Gemini 1.5 Pro001· LMArena data1,270
286Gemma 2 9B IT SimPO· LMArena data1,268.4
287Command R+08-2024· LMArena data1,268.3
288Claude 3.5 Sonnet20240620· LMArena data1,265.6
289GPT-4 Turbo2024-04-09· LMArena data1,264.8
290Claude 3 Opus20240229· LMArena data1,264.5
291Jamba 1.5 Large· LMArena data1,264.1
292GPT-4 Turbo Preview0125· LMArena data1,260.8
293Claude Haiku 3.520241022· LMArena data1,258.7
294Qwen 2.5 Coder 32B Instruct· LMArena data1,256.6
295Llama 3.1 Nemotron 51B Instruct· LMArena data1,252.1
296GPT-4 Turbo Preview1106· LMArena data1,249.2
297Aya Expanse 32B· LMArena data1,247.6
298Gemini 1.5 Flash-8B001· LMArena data1,247.5
299Gemini 1.5 Flash001· LMArena data1,247.4
300Nemotron-4 340B Instruct· LMArena data1,245.5

Showing 251–300 of 394 results.

Data credits: LMArena, CC BY 4.0 (licence)