SGLang describes running Muse Spark Open, which it calls the most capable US open-weight model to date, with NVFP4 MegaMoE, sequence parallelism, a DFlash2 draft model, and fine-tuning through Miles.
The post cites Meta's model specs as 816B total and 42B active parameters, a 1M token context window, and native text, image, video, and audio support. It credits Meta, NVIDIA, AMD, Intel, and radixark for the work.
The post explains how the SGLang stack serves Muse Spark Open for agentic work, adding speed and fine-tuning options to the model's release.
Muse Spark Open is built to plan, act, and use computer across agentic tasks for life and work.
You can now build a powerful personal AI agent with an open stack, and SGLang to make it:
We’re incredibly proud to have made this possible with @AIatMeta, @NVIDIAAI, @AIatAMD, @intel, and @radixark!
Today we’re releasing Muse Spark Open, the most capable US open-weight model to date. Built for coding and agentic work, it excels at long-horizon tasks, complex software engineering, and multi-step reasoning. Key Model Specs: → 816B total / 42B active parameters → 1M token context window → Natively multimodal across text, image, video, audio
Source: SGLang · x.comPublished · added here
