Muse Spark
Meta’s closed Muse Spark 1.3 on the Meta Model API and in Muse Code. Muse Wonder (2026-10-12) is the newer, higher-priced peak. Muse Spark Open (2026-10-15) is a separate weight release.
Meta Superintelligence Labs’ Muse Spark 1.3 — closed multimodal model for agentic tasks, long-horizon coding, computer use, and 1M-context workflows via Muse Code and the Meta Model API. As of 2026-10-12 it is the lower-priced API under Muse Wonder (muse-wonder-1.0), not Meta’s peak. Succeeds Spark 1.2; max reasoning is still withheld for safety testing. Muse Spark Open (2026-10-15) is a separate weight release, not this API id.
Why Muse Spark matters
Spark 1.3 remains Meta’s lower-priced paid API for agentic coding, computer use, and multimodal agents—distinct from open-weight Llama, from Muse Glimmer (Apache-2.0 30B on-device), and from Muse Wonder. Max reasoning is still withheld for safety testing. The Spark 1.3 API weights stay closed. Muse Spark Open (2026-10-15) is a separate 816B/42B-active multimodal checkpoint under the Muse Model License with limited commercial terms. Contributor mode is available for Spark, not for Wonder.
Vision · Audio input · Tool calling · Thinking · MCP · Coding
Last reviewed: 18 October 2026
When to choose Muse Spark
Decision guidance for architects—not a feature list.
Best for
- Long-horizon coding with Muse Code at a lower token price
- Personal agentic orchestration
- Computer-use workflows
- Multimodal perception + action loops
Avoid if
- You need the Spark 1.3 API weights themselves. Muse Spark Open is a separate checkpoint under the Muse Model License, which has limited commercial terms
- You need a mature terminal-agent ecosystem beyond Muse Code beta
Strengths
DataAIHub qualitative scores from 1 to 5, for architects. Not official benchmark results or a public ranking.
- Agentic coding / tools★★★★★
- Multimodal + computer use★★★★★
- API price competitiveness★★★★☆
- Open weights★☆☆☆☆
- Ecosystem maturity★★★☆☆
Ecosystem
Built by
- Meta
Muse Spark 1.3 is the lower-priced closed API under Muse Wonder.
Competes with
- GPT-5.6
Competes on agentic coding and multimodal tool use via a paid API.
- Claude Opus
Competes for long-running agents and computer-use style workflows.
Works with
Recommended for
- AI Agents
Built for multi-agent orchestration, tools/MCP, and computer use.
- Llama Models
Clarify open Llama vs closed Muse Wonder and Muse Spark when evaluating Meta stacks.
Often paired with
- Llama 4
Llama stays Meta’s open-weight generalist path; Muse Spark is the lower-cost closed API under Muse Wonder.
- Muse Glimmer
Glimmer is the Apache-2.0 on-device Muse sibling; Spark is the lower-cost 1M-context API under Muse Wonder.
How Muse Spark evolved
Key moments in chronological order.
- Open source
Meta releases Muse Spark Open, a separate weight checkpoint from the closed Spark 1.3 API: 816B total parameters, 42B active, 1M context, natively multimodal. License is the Muse Model License with limited commercial terms. Weights are on Hugging Face, and hosted inference is available from partner clouds. The Meta Model API Spark 1.3 id stays closed.
- Model
Meta releases Muse Wonder (muse-wonder-1.0) above Spark on the Meta Model API and in Muse Code. Spark 1.3 stays available at the lower token price, including Contributor mode. Wonder has no Contributor mode. Standard Wonder price is $4 / $0.20 cached / $20 per 1M tokens.
- Model
Improves long-horizon agentic work and coding in Muse Code and Meta Model API. Asks clarifying questions, confirms consequential actions, and uses fewer turns/tokens than 1.2 per Meta. Previously available reasoning modes are live; max reasoning is still withheld for safety testing. This API release did not include weights.
- Open source
Apache-2.0 30B on-device agent model (sibling to closed Muse Spark); Hugging Face meta-models/Muse-Glimmer-30B.
- Model
Coding-focused upgrade with more coding compute and environment diversity; available via Meta Model API with expanded global access.
- Product
Terminal coding agent powered by Muse Spark 1.2 with persistent background subagents and a replay-exact event-log runtime.
- Model
Muse Spark 1.1
Major gains in agentic coding, tool/computer use, and multimodal understanding.
- API
Meta Model API public preview
Meta’s first paid model API (OpenAI-compatible) opens Muse Spark to developers.
- Product
Thinking mode in Meta AI
Consumer Meta AI / meta.ai surfaces expose Muse Spark Thinking mode.
- Model
Muse Spark first release
Meta Superintelligence Labs ships its first closed multimodal Muse model.
Overview
Capabilities
- Vision: Yes
- Audio input: Yes
- Tool calling: Yes
- Thinking: Yes
- MCP: Yes
- Coding: Yes
- Structured output: Yes
Technical specifications
- Provider
- Meta
- License
- Proprietary
- Context window
- 1M
- Parameters
- Undisclosed
- Architecture
- Transformer (proprietary)
- Release
- 2026-09
- Modalities
- Text, Image, Video, Audio
- Vision
- Yes
- Audio input
- Yes
- Tool calling
- Yes
- Thinking
- Yes
- MCP
- Yes
- Open weights
- No
- API
- Yes
- Pricing (input)
- ~$1.25 per 1M tokens (Meta Model API / Muse Code standard)
- Pricing (output)
- ~$4.25 per 1M tokens
Supported modalities
Text · Image · Video · Audio
Context window
1M (1,000,000 tokens)
Pricing
Contributor tier discounts may apply when sharing data for training. Verify developer.meta.com / Muse Code pricing.
Availability
Muse Spark 1.3 (2026-09-02) remains in Muse Code and the Meta Model API at $1.25 / $0.15 cached / $4.25 per 1M tokens, including a Contributor-mode variant. Muse Wonder (muse-wonder-1.0, 2026-10-12) is the newer peak and has no Contributor mode. Previously available Spark reasoning modes ship now; max reasoning is withheld pending additional safety testing. Muse Spark Open (2026-10-15) is a separate checkpoint: 816B total / 42B active, 1M context, multimodal, under the Muse Model License with limited commercial terms. It is not the closed Spark 1.3 API.
Use cases
- Long-horizon agentic coding with Muse Code
- Personal agentic orchestration across apps
- Computer-use workflows
- Multimodal perception + action loops
Strengths
- 1.3 long-horizon agentic/coding upgrade vs 1.2 (fewer turns/tokens per Meta)
- 1M context with context management
- Multimodal + computer-use in one closed stack
Limitations
- The Meta Model API Spark 1.3 weights stay closed. Muse Spark Open is a separate checkpoint under the Muse Model License, which has limited commercial terms
- Newer ecosystem than OpenAI/Anthropic
- Muse Code still in beta
Related guides
Related benchmarks
Related research
- Introducing Muse Spark Open
Original Paper
- Introducing Muse Spark 1.3
Original Paper
- Introducing Muse Code and Muse Spark 1.2
Original Paper
- Introducing Muse Spark 1.1
Original Paper
Related tools
Related rankings
Related comparisons
Companies
Explore more models
- Muse WonderMetaMeta Superintelligence Labs’ Muse Wonder (API id muse-wonder-1.0, 2026-10-12) — closed model for longer agentic coding and professional tasks on the Meta Model API and in Muse Code. It sits above Muse Spark 1.3, which remains the lower-priced API. Muse Wonder Full is a Muse Code subscriber variant; the public pricing page lists only muse-wonder-1.0. Preview. Closed weights.
- Muse GlimmerMetaMeta Superintelligence Labs’ Muse Glimmer — Apache-2.0 ~30B dense multimodal agent model for on-device and single-GPU local agents. Sibling to closed Muse Spark and Muse Wonder; distinct from Llama 4.
- Llama 4MetaMeta’s Llama 4 family — open-weight multimodal models designed for research and commercial use under Meta’s community license.
- Claude FableAnthropicAnthropic’s Claude Fable 5.1 — the most capable widely released Claude for long-horizon agents, deep reasoning, and demanding coding workflows. Mythos 5.1 is the same underlying model with relaxed cyber/life-sciences safeguards for trusted-access programs.
- Claude OpusAnthropicAnthropic’s Claude Opus 5.5 (API id claude-opus-5-5, released 2026-09-22) for complex agentic coding, enterprise work, long-context analysis, and careful instruction following. Anthropic says it matches Fable 5.1 on most work at a lower price. Fable 5.1 remains a separate higher list-price SKU. Sonnet 5.5 (claude-sonnet-5-5, 2026-09-28) is the current Sonnet at the same $2/$10 list price as Sonnet 5. Haiku 5.5 (claude-haiku-5-5, 2026-10-07) is the current Haiku.
- Claude SonnetAnthropicAnthropic’s Claude Sonnet 5.5 (API id claude-sonnet-5-5, released 2026-09-28) for well-scoped everyday coding, agents, and documents. List price matches Sonnet 5 at $2/$10 per 1M tokens. Opus 5.5 remains the high-stakes tier; Haiku 5.5 (2026-10-07) is the low-cost subagent and high-volume tier.