GLM-5

[voipmonitor]

Hello,

do you eve plan to release dflash for GLM-5? If so and for FP8 - will it work for other quants like nvfp4?

[jianc99]

Hi, we are currently working on supporting Kimi-K2.5. Will support GLM 5 after that. For quantized target model, I think yes it works for various quants. We previously tested the Qwen3-8B-DFlash trained for the BF16 target model, and found that it works well for FP8 and NVFP4 version Qwen3-8B as well.