Muse Spark 1.1 is Meta AI's newest frontier model and its first paid developer API, launched in July 2026. It targets builders and enterprises who want a long-context, agentic coding and research model that can run subagents, control a computer across desktop, browser, and mobile, and tackle benchmark-hard tasks.
Core Features
- A 1,000,000-token context window that lets the model hold entire codebases or long documents in memory during a task.
- Native computer use across desktop, browser, and mobile, so it can click, scroll, and operate real applications.
- Parallel subagent delegation, letting the model spin up child agents to work on parts of a job at the same time.
- A public-preview paid API priced at $1.25 per million input tokens and $4.25 per million output tokens, with $20 in free credits at launch (US-only).
- Strong agentic benchmark results - number one on MCP Atlas, JobBench, and Humanity's Last Exam, and 20 percent on the held-back Vals AI Harvey legal-agent benchmark versus 11 percent for Fable.
Use Cases
- Autonomous coding agents that plan and edit across large repositories, with Replit, Cline, and Box named as early partners.
- Research and analysis workflows that need to reason over huge documents without chunking.
- Enterprises building computer-use assistants that operate software through a real interface.
Pricing
Free preview credits of $20 (US-only at launch). Paid API at $1.25 per million input and $4.25 per million output tokens. No open weights are released.
Our Take
A serious GPT-class challenger with best-in-class agentic scores, but the API is US-only in preview and closed-weight. Compare it with other engines in our AI model directory at aifreetool.site/tool-category/ai-engine-model/ and chat tools at aifreetool.site/tool-category/ai-chat-tools/.
