Meta launched Muse Wonder and Muse Wonder Full, its most capable AI models to date, for agentic tasks, coding, and long-running projects.
On October 12, 2026, Meta introduced Muse Wonder and Muse Wonder Full, calling them its most capable models to date. Built for agentic work, coding and extended projects, both offer a 1-million-token context window and support text, image, video and audio input, with text output. Meta says Active Memory carries useful context between conversations, while Active Focus helps models persist through obstacles and check with users before consequential actions. Muse Wonder launched through Meta’s API and Muse Code at $4 per million input tokens and $20 per million output tokens; Wonder Full is restricted to Muse Code subscribers. Meta says Wonder Full led selected coding evaluations, including SWE-Marathon, but comparisons were not controlled. The launch raises Meta’s challenge to OpenAI and Anthropic in paid developer tools and AI agents.