Reviewed by Jonathan West · Updated Jul 27, 2026

Muse Spark 1.1 Pricing and Access FAQ

A straightforward guide to Muse Spark 1.1 pricing, free-tier access, cost comparisons, and when preview pricing will change.

Reviewed by Jonathan West · Updated Jul 27, 2026

Status as of July 2026: Muse Spark 1.1 launched July 9, 2026, in public preview for US developers on the Meta Model API. This is Meta's first frontier model available to developers via paid API access (Meta).

Muse Spark 1.1 is a multimodal reasoning model built for agentic tasks, with strong performance on tool use, computer use, coding, and multimodal understanding. It ships at $1.25 per million input tokens and $4.25 per million output tokens—roughly one-quarter of OpenAI and Anthropic's flagship rates (Meta).

This guide answers the key pricing and access questions: Is there a free tier? What does it cost? When will preview pricing change? How does it compare to Claude Fable 5 and GPT-5.6? And which tier fits your use case? We use Meta's official announcement and developer documentation for every claim.

Is Muse Spark 1.1 Free? (Free Credits and Trial)

Muse Spark 1.1 is not fully free, but Meta gives every new developer $20 in free credits at signup on the Meta Model API (Meta). That covers initial exploration and testing at no cost.

The free credits are one-time only and consume quickly on heavy use. For sustained work, you move to pay-as-you-go after credits run out.

Public preview is available now in the US. If you need to try Muse Spark 1.1 before spending, the $20 credits are designed for pilots and proof-of-concept work.

  • Free tier: $20 free credits per new account on signup (Meta).
  • After credits: pay-as-you-go pricing at $1.25 input / $4.25 output per million tokens.
  • Availability: public preview for US developers via Meta Model API (Meta).
  • Timeline: available now as of July 9, 2026 (Meta).
  • Who pays after credits: any sustained or production use.
The $20 free credit is one-time and best used for small pilots. Plan for production costs if you move beyond testing.
Sources
Meta — Introducing Muse Spark 1.1

Not sure if Muse Spark 1.1 fits your cost profile and workflows, or how to roll out agentic models across your team? We can audit your token usage and map the right model to each task.

Book a Consultation

Pricing Model: How Muse Spark 1.1 Costs Break Down

Muse Spark 1.1 pricing is per-token: $1.25 per million input tokens and $4.25 per million output tokens (Meta). Input is what you send to the model. Output is what the model generates.

Output tokens cost 3.4x more than input tokens, which is typical for reasoning models. This means long, verbose responses drive the bill faster than long input prompts.

Meta set this pricing to undercut OpenAI and Anthropic. The goal is aggressive pricing that wins on cost while maintaining developer access (Meta announcements).

  • Input: $1.25 per million input tokens (Meta).
  • Output: $4.25 per million output tokens (Meta).
  • Token ratio: output costs 3.4x more than input.
  • Best-case cost: a 1,000-token input + 500-token output response costs roughly $0.004 (less than one-tenth of a cent).
  • Worst-case cost: a 100,000-token input + 50,000-token output response costs roughly $0.38.
  • Free tier: $20 credits at signup covers small pilots.
Output tokens cost 3.4x more than input. Long, verbose answers drive the bill. Cap output length where you can to control costs.
Sources
Meta — Pricing $1.25/$4.25 per M tokens

Muse Spark 1.1 Pricing vs Claude Fable 5 and GPT-5.6

Muse Spark 1.1 is the cheapest frontier model on the market today. Meta positioned it as one-quarter the cost of OpenAI and Anthropic flagship models (Meta). Here is the actual pricing comparison:

Claude Fable 5 costs $10 input / $50 output per million tokens (Anthropic). GPT-5.6 ranges from $1–$5 input / $6–$30 output depending on tier (OpenAI). Muse Spark 1.1 is $1.25 input / $4.25 output (Meta).

The practical message: Muse Spark 1.1 is dramatically cheaper per token than Claude Fable 5 and competitive or cheaper than even GPT-5.6's Luna tier (OpenAI's cheapest). If cost is the main driver, Muse Spark wins.

  • Muse Spark 1.1: $1.25 input / $4.25 output (Meta).
  • Claude Fable 5: $10 input / $50 output (Anthropic).
  • GPT-5.6 Sol (most expensive): $5 input / $30 output (OpenAI).
  • GPT-5.6 Luna (cheapest): $0.20 input / $1.20 output (OpenAI).
  • Verdict on cost: Muse Spark 1.1 is cheaper than Fable 5 on every token; competitive with or cheaper than GPT-5.6 depending on tier.
See our Muse Spark 1.1 vs Claude Fable 5 comparison and Muse Spark 1.1 vs GPT-5.6 comparison for a fuller cost breakdown with real-world scenarios.
Sources
Anthropic — Claude Fable 5 PricingOpenAI — GPT-5.6 Pricing

How to Access Muse Spark 1.1 Today

Muse Spark 1.1 is available now via the Meta Model API in public preview, US-only (Meta). Signup is self-serve with no vetting or waitlist.

You access it through the Meta Developer Console. Meta provides OpenAI-compatible and Anthropic-compatible request formats, so you can use familiar SDKs (Meta).

Getting started is straightforward: sign up, claim your $20 free credits, then integrate via the Meta Model API using standard HTTP calls or SDKs.

  • Where: Meta Model API (public preview, US-only).
  • Signup: self-serve, no vetting, no waitlist.
  • Free credits: $20 per new account at signup (Meta).
  • SDKs: OpenAI-compatible and Anthropic-compatible formats supported.
  • Status: public preview as of July 9, 2026 (Meta).
  • Access surface: REST API, Python SDK, web console.
Sources
Meta — Introducing Muse Spark 1.1

Access Tiers: Public Preview to Paid API to Enterprise

Muse Spark 1.1 ships in one tier today: public preview API access with pay-as-you-go pricing (Meta). There is no separate enterprise tier announced yet, and no tiered access (as of July 2026).

The path is simple: sign up for public preview, use $20 free credits, then pay per token once credits run out. Enterprise arrangements are likely coming but are not yet public.

Unlike some models, Muse Spark 1.1 does not gate you behind a waitlist or vetting. Self-serve access is the design.

  • Public preview: available now, US-only, self-serve signup.
  • Pricing: $1.25 input / $4.25 output per million tokens after free credits (Meta).
  • Enterprise: not yet announced as of July 2026.
  • Waitlist: none — direct signup available.

Real-World Cost Scenarios: Muse Spark 1.1 vs Competitors

Token math is hard to visualize. Here are three real scenarios to show what Muse Spark 1.1 costs versus Claude Fable 5 and GPT-5.6 (Luna tier, OpenAI's cheapest).

Scenario 1 (Light Use): 10 API calls, 5K input tokens + 500 output tokens per call. Total 50K input, 5K output. Muse Spark 1.1 = $0.08. Claude Fable 5 = $0.52. GPT-5.6 Luna = $0.06.

Scenario 2 (Document Coding Task): 100K input tokens (code review) + 20K output tokens (rewrite). Muse Spark 1.1 = $0.21. Claude Fable 5 = $1.50. GPT-5.6 Luna = $0.22.

Scenario 3 (Heavy Daily Use): 500K input tokens + 100K output tokens per day, 30 days. Total 15M input, 3M output. Muse Spark 1.1 = $30. Claude Fable 5 = $200. GPT-5.6 Luna = $12.

  • Scenario 1 (10 small API calls): Muse Spark wins, but GPT-5.6 Luna is close.
  • Scenario 2 (One large coding task): Muse Spark is cheapest, Claude 2.4x more expensive.
  • Scenario 3 (Heavy production use): GPT-5.6 Luna is 40% cheaper than Muse Spark 1.1, but Muse Spark beats Claude 6.7x on cost.
  • Takeaway: Muse Spark 1.1 is cheaper than Claude Fable 5 at all scales; sometimes cheaper than GPT-5.6 depending on your output ratio.
Token count matters more than the model. A thousand short API calls cost less than one long document review. Estimate your token volume before picking a model.

When Will Muse Spark 1.1 Pricing Change?

Muse Spark 1.1 pricing is currently in public preview, which typically signals a stability window (Meta). Preview pricing usually stays stable for 60–90 days or longer, though Meta has not published a specific end date.

The $20 free credit offer is tied to preview status. When the model exits preview, Meta may change pricing or retire the free credits.

Safe assumption: plan for pricing to stay stable through at least October 2026, then monitor Meta announcements for changes when the preview ends.

  • Preview status: launched July 9, 2026 (Meta).
  • Pricing lock: typically stable for 60–90 days in preview.
  • Free credits ($20): tied to preview; may end at GA.
  • When to watch: likely October–November 2026 for any GA and pricing changes.
  • Safe plan: lock in discounts and use credits while available.

Frequently Asked Questions

  • No, Muse Spark 1.1 is not free. However, Meta gives every new developer $20 in free credits at signup on the Meta Model API (Meta). After credits, you pay $1.25 per million input tokens and $4.25 per million output tokens.
  • Muse Spark 1.1 costs $1.25 per million input tokens and $4.25 per million output tokens (Meta). That is roughly one-quarter the cost of Claude Fable 5 and competitive with or cheaper than GPT-5.6 depending on tier.
  • The $20 free credit is a one-time signup offer on the Meta Model API. It covers initial exploration and testing. After it runs out, you move to pay-as-you-go pricing (Meta).
  • Yes. Muse Spark 1.1 costs $1.25/$4.25 per million tokens (Meta), while Claude Fable 5 costs $10/$50 (Anthropic). Muse Spark is roughly 88% cheaper on input and 91% cheaper on output tokens.
  • Compared to GPT-5.6 Luna (OpenAI's cheapest tier at $0.20/$1.20), Muse Spark 1.1 is 25% cheaper on input but 29% more expensive on output. For GPT-5.6 Terra or Sol, Muse Spark is cheaper on both input and output.
  • Muse Spark 1.1 is currently in public preview as of July 2026 (Meta). Preview pricing typically lasts 60–90 days, but Meta has not announced an exact end date. Monitor Meta announcements for changes when the preview ends.
  • Yes, for pilot and test work. The $20 free credit covers small pilots. For production use, you will pay per token after credits run out (Meta).
  • Not yet. As of July 2026, Muse Spark 1.1 ships via public preview API with pay-as-you-go pricing only (Meta). Enterprise pricing may come later.
  • Sign up for the Meta Model API (public preview, US-only). The signup is self-serve with no waitlist or vetting (Meta). You will get $20 free credits at signup and can start building immediately.
  • There is one public tier: the API model at $1.25/$4.25 per million tokens (Meta). Use it for all tasks. If cost is the main concern, keep output length short to reduce per-call expenses.

Contact Us