Meta opens 816B-parameter Muse Spark Open for coding agents

Meta's decision to release open weights for Muse Spark Open marks a shift for a model that had been proprietary since its April debut. The company built it on Muse Spark 1.2, shipped in August, and then added safety post-training so the model holds up without the guardrails of Meta's own services. Meta says the release cleared the thresholds in its Meta Superintelligence Scaling Framework, with a full evaluation report still to come.

The company is positioning the release carefully. The weights come with a new Muse Model License carrying "limited commercial terms," and while no fee applies to developers who adapt the model or ship it inside their own software, Meta requires a separate agreement from any company that sells the model as a hosted service and has passed $20 million in gross revenue over a 12-month period. Products built on it must carry a "Built with Muse Spark Open" label, and military and critical infrastructure uses are off-limits. Meta's previous open-weight release was the much smaller Muse Glimmer, a 30 billion-parameter model that went out in August, and the company's most capable systems remain closed.

The benchmark picture is mixed. In Meta's own table, Muse Spark Open scored 61.8 on the JobBench agent test, more than twice the 31.1 recorded by Inkling, the U.S. open model from Thinking Machines Lab, while Nvidia's Nemotron 3 Ultra trailed at 27.2. But Chinese open models are stronger at coding, with Kimi K3 leading the Terminal-Bench 2.1 coding test at 85 and GLM 5.3 and Qwen3.8 both finishing above Muse Spark Open's 76.4. On tool use, the order flips: Muse Spark Open's 89.2 on MCP-Atlas beat every rival Meta listed, the Chinese models included.