2.8T LatentMoE + KDA, refusal-surface reduced at the weight level. GGUF variants need llama.cpp PR #26185 — mainline cannot load them.
-
BlackfrostAI/KIMI-K3-DERISKED-MXFP4
Text Generation • 2.8T • Updated • 1 -
BlackfrostAI/KIMI-K3-DERISKED-MXFP4-GGUF
Text Generation • 2.8T • Updated -
BlackfrostAI/KIMI-K3-DERISKED-Q2_K-GGUF
Text Generation • 2.8T • Updated -
BlackfrostAI/KIMI-K3-DERISKED-CODER-REAP-320-MXFP4
Text Generation • 1T • Updated