Scale Agentic Coding Workflows with Meta’s Muse Spark Open on NVIDIA GB300 NVL72

Gives deployment-relevant specifics for running an 816B MoE coding model: top-8 of 256 expert routing, NVFP4 weights, and vendor-reported throughput up to 50K tokens/s per GPU. Useful for sizing self-hosted agent serving.

• Muse Spark Open has a 1M token context window, which the post says can hold an entire project with dependencies and thousands of files without splitting the task.
• The model uses goal conditioning and context compaction to stay on track in long sessions, and it writes separate commentary before each action so developers can see what the agent is doing.
• NVIDIA says the GB300 NVL72 links 72 Blackwell Ultra GPUs in one NVLink domain at 130 TB/s, which matters because routing each token to 8 of 256 experts generates heavy expert traffic.
• Dynamo recipes cover SGLang, vLLM, and TensorRT LLM, splitting prefill and decode onto separate GPU pools and routing requests to workers that already hold the relevant KV cache prefix.
• Fine-tuning is supported through NeMo AutoModel for full SFT and LoRA directly on Hugging Face checkpoints, with Megatron Bridge converting to Megatron format for larger production training.

• mixture-of-experts — A model built from many specialist sub-networks where only a few activate per token, giving big-model capability at small-model running cost.
• context compaction — Summarizing an agent's earlier conversation to free room in the context window so a long session can keep going.
• KV cache — The memory a model keeps about text it has already read, so generating each new token doesn't require reprocessing the whole conversation.
• LoRA — A cheap way to fine-tune a model by training a small add-on layer instead of changing all of the model's weights.

• articleOptimizing Agent Reliability and Tool Calling Efficiency for Chip Design Using NVIDIA NeMo Gym
• articleNVIDIA Run:ai Is Now Open Source as Part of NVIDIA DSX AI Factory Platform
• articleBuild Applications on NVIDIA BlueField Faster with NVIDIA DOCA Agent Skills
• articleAI Native by Design: Lessons Learned from Building NVIDIA TensorRT Model Connect

• postMeta ships Muse Spark Open, an 816B open-weight model for agentic codingAIatMeta
• articleMuse Spark Openbigwheels
• articleWelcome Muse Spark Open: open-weight frontier multimodal mixture-of-experts model from Metahuggingface.co

Checking sign-in…

Loading comments…