Muse Spark Open on Fireworks: Meta’s open weights wodel, available on US serverless

Fireworks announced day-one serverless availability of Meta’s Muse Spark Open, an 816-billion-parameter mixture-of-experts open-weights model that activates 42 billion parameters per token and supports a one-million-token context window for extended coding and agent workflows. The company positions the US-developed model as particularly suited to regulated legal, financial, and clinical applications, citing results from Vals AI, Rogo, Artificial Analysis, Meta, and its own validation suite that report strong performance in legal-agent tasks, finance workflows, clinical documentation, coding, reasoning, and image understanding, alongside a comparatively lower hallucination rate. Fireworks says the model runs on US-only capacity with zero data retention by default and no additional US pricing premium, and can be accessed through its serverless API, reserved throughput options, and LoRA-based serverless training for customization.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.