OsaurusAI/gemma-4-26B-A4B-it-qat-MXFP4

Quantization vs google/gemma-4-26B-A4B-it-qat-q4_0-unquantized. OsaurusAI/gemma-4-26B-A4B-it-qat-MXFP4 is a quantized version of the other (U32/F16).

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

On the Hub · Downloads on Model Pulse