Meta CEO Mark Zuckerberg returned to posting on X for the first time in three years to announce its latest AI model Muse Spark 1.1. Zuckerberg said in his post that X Muse Spark 1.1 is designed for agentic AI tasks, with improved coding, tool use, computer-use capabilities and multimodal reasoning.
“It does well on long-running tasks with 1M token context window, can delegate execution to sub-agents running in parallel, and is trained to use computer interfaces on desktop, mobile, or browser,” he added, and also said there is more coming soon.
READ: What is Meta Muse Image? Meta’s new AI image generator explained (July 8, 2026)
Meta also mentioned Muse Spark 1.1. is the first major release from its Meta Superintelligence Lab. The company said Muse Spark 1.1 is built to coordinate multiple AI agents working in parallel, and that the model l is designed to manage long-running tasks, retain context across extended sessions and determine when to automate workflows or interact directly with computer interfaces.
Meta also said the launch builds on this week’s release of Muse Image and reflects its broader push toward developing AI systems capable of reasoning, taking actions and assisting users across coding, productivity and multimodal tasks.
“If Muse Spark 1.1 is genuinely competitive with Claude and Codex on coding, then Meta may finally have a much clearer monetization bridge from AI models to paid developer tools,” said Shay Boloor, chief market strategist at Futurum Equities.
Developers in the United States can now access Muse Spark in public preview on Meta Model API, letting them test prompts, compare outputs and prototype integrations. People who sign up for the API receive $20 in free credits to test the model before switching to pay-as-you-go pricing.
Zuckerberg also said on X, “Our focus is on delivering strong agentic and multimodal models at very low cost.”
The access is priced at $1.25 per million input tokens and $4.25 per million output tokens. This comes above OpenAI’s entry-level GPT‑5 mini and Anthropic’s low-cost Claude Haiku 4.5, but below Anthropic’s higher-end Claude Sonnet 4.6 model.
READ: Meta faces possible action in India over ads promoting child abuse (July 6, 2026)
The model is now available in Thinking mode in the Meta AI app and on the website. It is expected that it would replace existing Llama models powering chatbots on WhatsApp, Instagram, Facebook and Meta’s collection of smart glasses.
The announcement about Muse Spark 1.1. comes shortly after Meta introduced Meta Muse Image, its first dedicated image-generation model developed by Meta Superintelligence Labs. The Meta Muse Image model is being integrated into Meta AI, the company’s chatbot, enabling users to create images from text prompts, edit existing photos, and generate visuals using sketches or annotations. The rollout also brings a range of AI-powered creative features to Instagram, WhatsApp, Facebook, and Messenger, with broader availability expected in the coming months.