LLM Benchmarks

Reference pages for the benchmarks the LLM community actually uses to compare models. Each benchmark page explains what's measured, how, and a leaderboard of all models in the database with scores on this benchmark.

All benchmarks

BenchmarkCategoryMetric
Aider Polyglot coding pass@1
HumanEval coding pass@1
LiveCodeBench coding pass@1
MBPP coding pass@1
SWE-bench Verified coding pass@1
Chatbot Arena general elo
MMLU general accuracy
MMLU-Pro general accuracy
TruthfulQA general accuracy
IFEval instruction-following accuracy
AIME 2025 math accuracy
MATH math accuracy
MMMU multimodal accuracy
GPQA Diamond reasoning accuracy
HellaSwag reasoning accuracy