Meta AI has released Muse Code (in beta), a terminal coding agent in beta, powered by its new Muse Spark 1.2 model. Meta positions the pair as its next step toward the frontier, with larger models on the way. Muse Code targets complex software engineering across large repositories: it plans changes, writes code, and validates the results. A set of async background agents stays alive for the whole session instead of spawning per task. A local append-only event log records every model call, tool run, approval, and edit, which Meta calls replay-exact and restart-safe. Muse Spark 1.2 was co-trained with the harness itself. Meta also published a kernel-optimization case study running 1,000+ tool calls over as long as 24 hours.
Is it deployable
Yes. Muse Code ships in beta for macOS and Linux via curl -fsSL https://dev.meta.ai/install.sh | bash. Muse Spark 1.2 is available in Muse Code and the Meta Model API, with expanded global access. The launch post does not mention downloadable weights, so treat this as a hosted dependency.
- Company level: The API path fits any size. The Muse Code path fits teams already running agents in sandboxes with review gates.
- Industries: Software and SaaS, developer tooling, fintech engineering, GPU and inference infrastructure, semiconductors and HPC.
- Applications: Repository-scale refactors and migrations, long-running bug triage, test generation, and GPU kernel optimization.
Async background agents
Muse Code runs a simple agent loop plus a set of async background agents. These specialized agents remain active throughout each session. They are not spawned for individual tasks, which Meta says avoids redundant information gathering. They carry out next steps and choose when to report back to the main agent. Meta states this persistence reduces latency and steering on difficult, multi-step tasks.
Runtime design
Muse Code uses a local event log. Every model call, tool run, approval, and edit is appended to it. Meta calls this single source of truth replay-exact and restart-safe. After a crash, the agent resumes precisely where it stopped, letting long-running tasks survive failures.
Bundled skills
Three default skills ship with the agent. /plan turns a task into an approval-gated plan. /grill stress-tests that plan until it holds up. /goal works toward successful completion of the specified objective.
What changed in Muse Spark 1.2
Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1. Meta reports gains in code generation, complex debugging, codebase understanding, and end-to-end developer workflows. The research team significantly scaled up training compute on coding tasks and expanded environment diversity. The model keeps its strength in other areas, including general agents.
Three important training details:
- Co-training with the harness: Muse Spark 1.2 was co-trained with Muse Code. Training included rejection-sampled harness trajectories and recipe optimizations for goals, compaction, and subagents. The Muse Code toolset was integrated to maximize harness compatibility.
- Long-horizon training: Training covered whole-repository generation, large end-to-end projects, and auto-research. The model uses planning, goal conditioning, and context compaction to sustain progress.
- Self-improvement: Muse Spark 1.1 generated challenging coding environments and instruction-following templates, then graded candidate solutions against those requirements. That produced a scalable training dataset for 1.2.
Evaluation
Meta’s methodology report is unusually specific. Terminal-Bench 2.1 uses all 89 tasks, pass@1 over five attempts. DeepSWE v1.1 covers 113 tasks across 91 repositories and five languages. Meta Internal Coding Bench holds 440 tasks derived from real internal pull requests. Runs execute in isolated Daytona cloud sandboxes. Comparisons include Grok 4.5, Claude Opus 5, GPT-5.6 Terra, Gemini 3.6 Flash, and Kimi K3, each with its own agent product. Meta notes its harness may not be tuned for third-party models. For reference, Meta’s model page lists Muse Spark 1.1 at 80.0 on Terminal-Bench 2.1.
