Meta prices Muse Spark 1.1 at $1.25/1M input tokens and $4.25/1M output tokens; Alexandr Wang says improving coding and agentic performance was a key focus
Ina Fried /Axios:
Context & Ripple Effects
Meta moved Muse Spark from a private API preview for select partners toward broader paid access, then launched the Meta Model API with pricing positioned at roughly a quarter of OpenAI and Anthropic alternatives. Publishing Muse Spark 1.1 token rates makes that positioning concrete.
The model effort sits within Meta Superintelligence Labs under Alexandr Wang, with Muse Spark also intended to improve Meta AI. The stated emphasis on coding and agentic work aligns the commercial API with capabilities Meta is also building into its products.
First-order effects
- Developers can evaluate Muse Spark 1.1 against competing APIs using explicit $1.25-per-million input-token and $4.25-per-million output-token rates, lowering the uncertainty around adopting Meta's model API.
- Meta gains a direct route to monetize Muse Spark outside its own products while using coding and agentic performance as its primary workload pitch.
Second-order effects
- Lower published token prices pressure other frontier-model API vendors to defend pricing or demonstrate enough performance, reliability, or tooling advantage to justify a premium.
- For application builders with code-generation or agent workflows, model selection becomes more sensitive to output-token costs, since those workloads can generate substantial model output.
Third-order effects
- If Meta sustains its stated low-price posture, frontier-model competition could shift further from model access alone toward the economics and developer experience of operating agentic applications.
- Meta's simultaneous use of Muse Spark in Meta AI and sale through an API points to a broader convergence between consumer AI distribution and enterprise developer platforms; whether it becomes a durable cloud-style business depends on adoption beyond Meta's own ecosystem.
The trend: This is one data point in the shift from closed frontier models as premium endpoints toward aggressively priced APIs competing for coding and agentic workloads.