InferenceIllusionist/gpt-oss-20b-MLX-4bit

Quantization vs openai/gpt-oss-20b. InferenceIllusionist/gpt-oss-20b-MLX-4bit is a quantized version of the other (U32/BF16).

5 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse