Open Models Almanac
Methodsoptimizationquantization-aware training

implementation detail · filed under optimization

MXFP4 Weights with MXFP8 Activations

A mixed-precision configuration using MXFP4 weights and MXFP8 activations; the evidence does not establish a distinct training algorithm.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See quantization-aware training for the methods filed beside it.

Also called MXFP4/MXFP8 quantization-aware training.

sources
2
model
1
lab adopt it
1
strongest
used

Documented in

Moonshot AI

Evidence

3 quotes from 2 sources, one card per model, each opening with its strongest.

How sources treat it
used 3

Kimi K3

Moonshot AIadopted2 sources

Kimi K3 applies quantization-aware training from the SFT stage onward, using MXFP4 weights with MXFP8 activations for broad hardware compatibility.

usedoptimizationin Kimi K3Moonshot AI
2 more from this model

Kimi K3 applies quantization-aware training from the SFT stage onward, using MXFP4 weights with MXFP8 activations for broad hardware compatibility.

usedpost trainingin Kimi K3Moonshot AI

MXFP4 weights / MXFP8 activations (quantization-aware training)

usedmodel architecturein Kimi K3Moonshot AI

Further reading

Picked by hand, not extracted: where to read more, not evidence for anything on this page.

↑