IFM/K2-Horizon-7B-Uno · Hugging Face - 5200tps with no quality loss

IFM released K2-Horizon-7B , diffusion augmented LLM at upto 5200 tps, with claimed lossless speedup. causal LLM architecture and adds a plug-and-play diffusion adapter alongside the autoregressive weights. https://arxiv.org/abs/2609.04010