Skip to content

Benchmark

Arena ELO - Instruction Following

Instruction following benchmarks measure how closely models follow instructions. Higher scores are better.

Models measured
366
Scores recorded
423
Sources
1
IQ weight
1.18

Score distribution

1,294.5
Mean
1,332.7
Median
155.5
Std dev
1,056.7
10th percentile
1,465
90th percentile
780.7
Min
780.71,539.1
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.18 for chat models, measured over 366 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
251GLM 4 Plus0111· LMArena data1,292.4
252Step-2 16K Exp202412· LMArena data1,286.7
253Qwen3 30B A3B· LMArena data1,283.1
254DeepSeek V2.51210· LMArena data1,279.5
255Hunyuan Turbo0110· LMArena data1,278.3
256Yi Lightning· LMArena data1,278.2
257GPT 4o2024-05-13· LMArena data1,277.5
258Hunyuan Large2025-02-10· LMArena data1,277
259Qwen2.5-Plus1127· LMArena data1,274.5
260Olmo 3 32B Think· LMArena data1,274.2
261Granite 4.2 3B· LMArena data1,272.5
262Athene V2 Chat· LMArena data1,270.5
263Molmo 2 8B· LMArena data1,269.9
264Grok 22024-08-13· LMArena data1,269.8
265Claude 3.5 Sonnet20240620· LMArena data1,269.5
266Granite 4.1 8B· LMArena data1,267.8
267Llama 4 Maverick· LMArena data1,266.9
268GPT 4o2024-08-06· LMArena data1,266.6
269GPT-4.1 nano2025-04-14· LMArena data1,266.3
270GLM 4 Plus· LMArena data1,264.2
271Mistral Small 3.1 24B Instruct2503· LMArena data1,262.8
272Qwen Max0919· LMArena data1,261.7
273Llama 3.1 405B InstructBF16· LMArena data1,258.5
274Llama 3.1 405B InstructFP8· LMArena data1,258.5
275GPT-4o mini2024-07-18· LMArena data1,258.3
276Gemini 1.5 Flash002· LMArena data1,257.4
277Gemma 3n E4B IT· LMArena data1,255.1
278Qwen2.5 72B Instruct· LMArena data1,253.9
279Gemini 1.5 Pro001· LMArena data1,253.5
280Magistral Medium2506· LMArena data1,253.4
281Gemini Advanced0514· LMArena data1,253.3
282Llama 3.1 Nemotron 70B Instruct· LMArena data1,251.7
283Hunyuan Large Vision· LMArena data1,250.3
284GPT-4 Turbo2024-04-09· LMArena data1,249.3
285DeepSeek V2.5· LMArena data1,249
286Mistral Large2411· LMArena data1,248.7
287Claude 3 Opus20240229· LMArena data1,248.2
288Mistral Large 22407· LMArena data1,247.7
289Llama 4 Scout· LMArena data1,247.5
290Olmo 3.1 32B Think· LMArena data1,246.5
291Hunyuan Standard2025-02-10· LMArena data1,244.8
292Grok 2 mini2024-08-13· LMArena data1,244.6
293Llama 3.3 70B Instruct· LMArena data1,241.7
294Claude Haiku 3.520241022· LMArena data1,240.5
295GPT-4 Turbo Preview1106· LMArena data1,240.5
296Gemma 3 4B IT· LMArena data1,238.8
297Mercury· LMArena data1,238.8
298GPT OSS 20b· LMArena data1,236.1
299Nova Pro 1.0· LMArena data1,235.2
300GPT-4 Turbo Preview0125· LMArena data1,234.9

Showing 251–300 of 423 results.

Data credits: LMArena, CC BY 4.0 (licence)