How much does Muse Spark 1.1 cost?

Muse Spark 1.1 costs $1.25 per million input tokens and $4.25 per million output tokens on the Meta Model API. New accounts get $20 in free credits at signup, and the launch preview is limited to US developers. Meta priced Muse Spark 1.1 below Anthropic's Claude and OpenAI's GPT to attract API workloads.

Is the Meta Model API OpenAI-compatible?

Yes, the Meta Model API is OpenAI-compatible. Developers can point existing OpenAI SDK code at Meta's endpoint and call Muse Spark 1.1 with parallel tool calling, structured output, and built-in search with citations. This lowers migration friction for teams already running on OpenAI or on Meta's early-partner integrations at Replit, Cline, and Box.

What can Muse Spark 1.1 read?

Muse Spark 1.1 handles images, video, audio, and PDFs inside a 1-million-token context window, so an agent can attach a long PDF and a screen recording in the same call. Meta says the model uses that context while operating tools and computers, so Muse Spark 1.1 targets multi-step agent workloads, not just chat.

How does Muse Spark 1.1 compare to the original Muse Spark?

The original Muse Spark shipped April 8 inside the Meta AI app but had no public API. Muse Spark 1.1 upgrades coding and computer-use performance, adds parallel tool calling and structured output, and — for the first time — opens the model behind Meta AI Thinking mode to outside developers through the paid Meta Model API.

Meta Superintelligence Labs · 2026-07-09 · major

Muse Spark 1.1 — Meta MSL opens its first paid API at $1.25 / $4.25 per million tokens

Muse Spark 1.1 is Meta Superintelligence Labs' first model available through a paid Meta Model API. The multimodal agentic model has a 1M-token context, an OpenAI-compatible endpoint, and undercuts Claude and GPT on price at $1.25 in and $4.25 out per million tokens.

Meta Muse Spark 1.1 announcement graphic from Neowin's coverage of the July 9, 2026 launch — Neowin / Meta

Meta finally opens Muse Spark to outside developers — with a price tag under Claude and GPT.

Key specs

Context window	1M tokens
Price (input)	$1.25 / 1M tokens
Price (output)	$4.25 / 1M tokens

Quick facts

Maker	Meta Superintelligence Labs
Context window	1 million tokens
Modalities	Text, images, video, PDFs, audio
Availability	Meta Model API (public preview, US), Meta AI Thinking mode, meta.ai
API surface	OpenAI-compatible
Launch credits	$20 free for new Meta Model API accounts
Early partners	Replit, Cline, Box

Pricing

Input	$1.25 / 1M tokens
Output	$4.25 / 1M tokens
Launch credit · free credits for new Meta Model API accounts	$20

source ↗

What is it?

Muse Spark 1.1 is a July 9 upgrade to Meta Superintelligence Labs' flagship model, plus the launch of the paid Meta Model API. The model is multimodal across text, images, video, PDFs, and audio, carries a 1-million-token context window, and now sits behind an OpenAI-compatible endpoint that developers can call directly. Replit, Cline, and Box are early API partners.

How does it work?

Behind the OpenAI-compatible surface, Muse Spark 1.1 exposes parallel tool calling, structured JSON output, and built-in web search with inline citations, so an agent can plan, look things up, and call tools in one round trip. The model runs in a Thinking mode that reasons through the 1M-token context while operating a computer or browser, and Meta positions the coding stack as competitive with leading alternatives on its Meta Internal Coding Bench.

Why does it matter?

Muse Spark went four months as a Meta-app-only model with no way for outsiders to build on it. The Meta Model API changes that: at $1.25 in and $4.25 out per million tokens, Muse Spark 1.1 lands below Anthropic and OpenAI's list prices, which reshapes the price floor for a 1M-context multimodal agent model. It is also Meta's first paid AI product, a shift from the open-weights-and-ads posture that defined Llama.

Who is it for?

Agent builders and coding tools that want a cheaper alternative to Claude or GPT with a real 1M-token context and native PDF, image, and video input.

Frequently asked questions

How much does Muse Spark 1.1 cost?: Muse Spark 1.1 costs $1.25 per million input tokens and $4.25 per million output tokens on the Meta Model API. New accounts get $20 in free credits at signup, and the launch preview is limited to US developers. Meta priced Muse Spark 1.1 below Anthropic's Claude and OpenAI's GPT to attract API workloads.
Is the Meta Model API OpenAI-compatible?: Yes, the Meta Model API is OpenAI-compatible. Developers can point existing OpenAI SDK code at Meta's endpoint and call Muse Spark 1.1 with parallel tool calling, structured output, and built-in search with citations. This lowers migration friction for teams already running on OpenAI or on Meta's early-partner integrations at Replit, Cline, and Box.
What can Muse Spark 1.1 read?: Muse Spark 1.1 handles images, video, audio, and PDFs inside a 1-million-token context window, so an agent can attach a long PDF and a screen recording in the same call. Meta says the model uses that context while operating tools and computers, so Muse Spark 1.1 targets multi-step agent workloads, not just chat.
How does Muse Spark 1.1 compare to the original Muse Spark?: The original Muse Spark shipped April 8 inside the Meta AI app but had no public API. Muse Spark 1.1 upgrades coding and computer-use performance, adds parallel tool calling and structured output, and — for the first time — opens the model behind Meta AI Thinking mode to outside developers through the paid Meta Model API.

Try it

muse-spark-1-1 on the Meta Model API (public preview, US developers).