Muse Spark 1.1
Meta's first paid API model — the agentic Spark upgrade with 1M context at $1.25/$4.25, public preview.
About
Muse Spark 1.1, released July 9, 2026 by Meta Superintelligence Labs, is the first paid model in the Muse family and the first Meta model sold through a commercial developer API. It upgrades the April Muse Spark with major gains in tool use, computer use, coding, and multimodal understanding.
It carries a 1 million token context window and is built for agentic work: native tools, MCP servers, custom skills, and multi-agent coordination where the model plans tasks, delegates to sub-agents, and combines outputs. It is available in Thinking mode through the Meta AI app and meta.ai, and via the new Meta Model API in public preview with an OpenAI-compatible interface.
Pricing is $1.25 per million input tokens and $4.25 output, roughly a quarter of comparable Anthropic and OpenAI rates, with $20 in free credits for new accounts. The API launched US-only on Meta's own properties, not third-party platforms, with new developers onboarded from a waitlist. Meta reports it leads Opus 4.8 and GPT-5.5 on MCP Atlas (88.1), JobBench (54.7), Humanity's Last Exam with tools, and Finance Agent v2, while trailing on SWE-Bench Pro, OSWorld-Verified, and Terminal-Bench 2.1.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported (Meta): MCP Atlas 88.1 vs Opus 4.8 82.2 vs GPT-5.5 xhigh 75.3; JobBench 54.7 vs 48.4 vs 38.3; leads on HLE with tools and Finance Agent v2; trails on Toolathlon-Verified, OSWorld-Verified, SWE-Bench Pro, Terminal-Bench 2.1. Independent: no published independent benchmark scores at research time.
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Release facts (July 9, 2026; Meta Superintelligence Labs; $1.25/$4.25 pricing; 1M context; Meta Model API public preview; $20 free credits; US-only launch) corroborated across Analytics India Magazine, WithO2, BizScoreAI, SEOScanPro, and Blockchain.News. Docked for: (1) benchmark figures are Meta's own comparison sheet with no independent verification; (2) public preview status means pricing and behavior may change; (3) proprietary with no weights and waitlist-gated access.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.0/10
- Maintenance · recency, releases, issue load
- 10.0/10
- Openness · license + docs
- 2.0/10
- Manual review · sanity read + red-flag check
- 8.0/10
- · Adoption 7: Meta Model API public preview; OpenAI-compatible endpoint; $20 free credits
- · Maintenance 10: released 2026-07-09; current Muse API model line (1.2 Aug, 1.3 Sep followed)
- · Openness 2: proprietary, no weights, waitlist-gated public preview
- · Manual 8: release facts (Jul 9 2026; $1.25/$4.25; 1M context; public preview API; $20 credits; US-only) corroborated across AIM, WithO2, BizScoreAI, SEOScanPro; docked for provider-only benchmarks, preview status, closed weights