Which AI Model Is
Actually the Best?

Real benchmark data from LMSYS Arena. Compare models by overall performance, category, or head-to-head.

Models

10+

Benchmarks

Arena ELO

Live Ranking

Jan 2025

Last Updated

Overall Model Rankings

Chatbot Arena ELO (human preference rating) and individual benchmark scores.

LMSYS Chatbot Arena ELO ScoreHigher is Better

#1GPT-4o1314

#2Claude 3.5 Sonnet1298

#3Gemini 1.5 Pro1261

#4Llama 3.1 405B1247

#5Grok 21235

#6Mistral Large 21219

Select a category to see which model performs best in that domain.

Coding Performance ScoreCoding

🏆Claude 3.5 Sonnet leads Coding with a score of 92

Select up to 4 models to compare across all dimensions.

Up to 4 models

Capability Radar

Category scores across 6 domains

Benchmark Comparison

MMLU · HumanEval · MATH · GSM8K · GPQA · BBH

Spec	GPT-4o	Claude 3.5 Sonnet	Gemini 1.5 Pro
Arena ELO	1314	1298	1261
Context Window	128K tokens	200K tokens	1000K tokens
Speed	110 t/s	85 t/s	150 t/s
Latency (P50)	890ms	1100ms	700ms
Input / 1M	$2.50	$3.00	$1.25
Output / 1M	$10.00	$15.00	$5.00
Free Tier	✓	✓	✓