Meta Unveils Muse Spark 1.1 and Model API, Marking Its First Paid AI Offering

After a three‑year silence on the platform formerly known as Twitter, Mark Zuckerberg used X to announce Muse Spark 1.1, Meta’s first agentic AI model offered through a paid developer API. The move places Meta squarely against OpenAI, Anthropic and Google, not just on technical capability but on price.

Muse Spark 1.1 is described as an “agentic” model – one that can plan tasks, invoke external tools, and operate a computer across desktop, mobile and browser environments. It supports up to one million tokens of context, allowing it to handle long‑form, multi‑step projects. In practice, the model can write and debug code, analyse images and video, and split work among parallel helper agents. The capability is now accessible in “Thinking” mode within the Meta AI app and on meta.ai, with a free $20 credit for new developers.

The pricing structure is the headline‑grabbing element. Meta lists $1.25 per million input tokens and $4.25 per million output tokens, undercutting Anthropic’s Claude Sonnet 5 ($3 and $15) and even its cheapest Claude Haiku 4.5. Elon Musk’s SpaceXAI released Grok 4.5 the same day at $2 per million input tokens, making Meta’s rate the most aggressive in the market.

Why does price matter? AI model usage is often measured in billions of tokens; a few dollars per million can translate into millions of dollars in annual spend for large‑scale applications. By positioning Muse Spark 1.1 as a lower‑cost alternative, Meta hopes to attract developers who might otherwise stay with OpenAI’s ChatGPT or Anthropic’s offerings, thereby building a revenue stream beyond advertising.

The launch also marks a strategic shift for Meta. Earlier this year the company distributed its Llama models for free to gain market share. The new Model API signals a move toward monetisation of core AI assets, a step that aligns with Meta’s expanded 2026 capital budget of up to $145 billion. The revenue potential is still modest – the API is a nascent product – but it diversifies the company’s income and ties it to the utilisation of its own AI‑focused data centres.

Investors reacted cautiously. META stock rose roughly 2% after the announcement, then steadied near $614, a muted response compared with the hype surrounding the technical details. Analysts point to Meta’s recent job cuts, a $14.3 billion acquisition of Scale AI, and the broader market’s fatigue with lofty AI promises as factors tempering enthusiasm.

From a broader industry perspective, Muse Spark 1.1 illustrates the accelerating commoditisation of agentic AI. Companies that once offered free, open‑weight models are now packaging similar capabilities behind usage fees. This trend could accelerate automation across software development, content creation and data analysis, as lower barriers to entry enable startups and mid‑size firms to embed sophisticated AI without prohibitive costs.

Real‑world implications are already emerging. Coding platforms such as Replit and Cline have signed on as early partners, planning to integrate Muse Spark 1.1 into their development environments. If the pricing holds, small teams could automate routine coding tasks, freeing engineers to focus on higher‑level design work. Likewise, media organisations could use the model’s multimodal abilities to generate video captions or edit footage, reshaping newsroom workflows.

In structural terms, the shift from free model distribution to a paid API reflects a maturing AI ecosystem where revenue generation, data centre utilisation and platform lock‑in become central considerations. Meta’s approach may prompt rivals to revisit their own pricing strategies, potentially igniting a broader price competition that could benefit developers but compress margins for AI labs.