MagicArena é uma plataforma de avaliação comparativa competitiva projetada para avaliar e classificar modelos visuais de IA generativa por meio de comparação lado a lado com humanos.

AGI-Eval é uma comunidade de avaliação especializada, criada para comparar as capacidades e o desempenho de diversos modelos de linguagem de IA de grande escala.

An advanced evaluation system by H2O.ai that utilizes Elo rating methodologies to benchmark and rank Large Language Models (LLMs).

A professional evaluation benchmark from Fudan University’s NLP Lab designed to measure the performance and reliability of large language models.

MMBench is a comprehensive evaluation framework designed to measure the capabilities of multimodal large language models across a wide array of visual and textual tasks.

A standardized, holistic evaluation framework from Stanford University designed to measure the performance and safety of large language models.

OpenCompass is an open-source evaluation framework developed by the Shanghai AI Lab to provide standardized, comprehensive benchmarking for large language models.

An open-source evaluation framework developed by the Beijing Academy of Artificial Intelligence (BAAI) to standardize and scale LLM benchmarking.

A crowdsourced benchmarking platform where users battle-test Large Language Models through blind side-by-side comparisons.

MMLU is a comprehensive benchmark designed to evaluate the general knowledge and problem-solving capabilities of large language models across a vast array of disciplines.