AI Briefing#int4
The latest AI and developer news about #int4, with the original source and a short summary.
The latest AI and developer news about #int4, with the original source and a short summary.
The INT4 quantized version of the NVIDIA Cosmos3 64B model has been released for CUDA and MLX, enabling local execution.
ICLR 2026's ParoQuant has released a 4-bit quantization method along with code and models for reasoning LLMs.
On 2 nodes with 32 AMD MI50s, Kimi K2.6 int4 achieved 9.7 tok/s output and 263 tok/s input.
The cause of KV cache INT4 collapse in some models was identified and fixed with a 4-line correction.