Furiosa-LLM (FLM) 2026.4.0 release and its Q4 roadmap

Hi folks, We released Furiosa SDK 2026.4.0 today! See the release announcement for details. Looking ahead, here are our priorities for Q4: Speech recognition models and Transcribe API support Video modality support KV cache management for linear and hybrid attention models HiCache general availability and further KV cache transfer optimizations DFlash speculative decoding support Expanded multi-token prediction (MTP) support, including Gemma 4 and Qwen3.5 Lower token generation latency through improved overlap scheduling Improved serving and KV cache observability Additional model support: Qwe