ERNIE 4.5 21B A3B MLX 8bit

ERNIE

Open WeightModel weights are publicly available — can be downloaded and self-hosted

Released July 2025131K context≈ 98,304 words

A mid-sized Chinese language model from Baidu's ERNIE family, quantized to 8-bit and optimized for Apple Silicon via the MLX framework. With 21 billion total parameters but only 3 billion active at a time through a mixture-of-experts architecture, it balances capability with efficiency. It handles a 131K token context window, making it capable of processing long documents on local hardware.

Capabilities

Capability scores are AI-generated based on model documentation, benchmarks, and technical specifications. Learn more

Factual Knowledge

Moderate

Tool Use

ERNIE 4.5 21B A3B MLX 8bit

ERNIE

Open WeightModel weights are publicly available — can be downloaded and self-hosted

Released July 2025131K context≈ 98,304 words

A mid-sized Chinese language model from Baidu's ERNIE family, quantized to 8-bit and optimized for Apple Silicon via the MLX framework. With 21 billion total parameters but only 3 billion active at a time through a mixture-of-experts architecture, it balances capability with efficiency. It handles a 131K token context window, making it capable of processing long documents on local hardware.

Capabilities

Capability scores are AI-generated based on model documentation, benchmarks, and technical specifications. Learn more

Factual Knowledge

Moderate

Tool Use

Glossary

Apple SiliconApple's custom-designed processors (like M1, M2, M3) optimized for running machine learning models on Mac computers.ArchitectureThe underlying structural design of a neural network that defines how data flows through layers and components.Context WindowThe maximum number of tokens a model can process in a single conversation or prompt.Language ModelAn AI model trained to predict and generate text by learning patterns from large amounts of written data.MLXA machine learning framework optimized for running models efficiently on Apple Silicon chips.MLX FrameworkA machine learning framework specifically designed for running AI models efficiently on Apple Silicon hardware.ParametersThe learned numerical values in a model — more parameters generally means more capacity but higher compute cost.QuantizedA technique that reduces a model's size and memory usage by storing weights with lower precision (fewer bits), trading some accuracy for efficiency.TokenA small unit of text (a word, subword, or punctuation mark) that a language model breaks input into for processing.

Capabilities

Capabilities

Use Case Fit

Glossary