Muse Spark Is Meta’s First Closed-Source Foundation Model
Meta Superintelligence Labs unveils Muse Spark, a natively multimodal model featuring advanced reasoning modes and 10x compute efficiency compared to Llama 4.
Meta introduced Muse Spark on April 8, 2026, marking a fundamental architectural departure from the Llama series. Developed by the newly formed Meta Superintelligence Labs, the model abandons the open-weights approach for a closed-source, proprietary API model. For developers, this signals a shift in baseline foundational models and introduces a native multi-agent orchestration layer directly within the model architecture.
Architecture and Reasoning Modes
Muse Spark is a natively multimodal foundation model. It processes visual and text inputs directly within its internal logic layer, enabling visual chain of thought for analyzing video frames and complex diagrams. The model utilizes a training technique Meta calls thought compression. This allows Muse Spark to match the performance of the prior Llama 4 Maverick while consuming ten times less compute.
The architecture introduces two distinct reasoning states. Thinking Mode activates deeper reasoning for complex single-domain tasks. Contemplating Mode functions as an internal orchestrator that spins up and manages multiple sub-agents in parallel. If you build multi-agent systems, this pushes the orchestration layer down into the model itself, handling concurrent agent execution natively rather than relying on external framework loops.
Benchmark Performance
Evaluations place Muse Spark squarely within the top tier of frontier models. The system is highly optimized for healthcare and scientific domains, achieving 38% on the FrontierScience Research benchmark after training on data curated by over 1,000 physicians.
On the Humanity’s Last Exam (HLE) evaluation, the base model scored 39.9%. When equipped with external tool access, this score increased to 50.4%. Artificial Analysis currently ranks Muse Spark tied for fourth place globally with an overall intelligence score of 52.
| Model | HLE Score (Base) |
|---|---|
| Gemini 3.1 Pro Preview | 44.7% |
| GPT-5.4 | 41.6% |
| Muse Spark | 39.9% |
Availability and Distribution
Muse Spark is currently available to consumers in the United States through the Meta AI application and web interface. The underlying infrastructure will deploy to WhatsApp, Instagram, Facebook, Messenger, and Ray-Ban Meta smart glasses in the coming weeks.
Developer access is restricted to a private API preview for select partners. This closed-source release breaks from the precedent set by the Llama series, removing a heavily relied-upon open standard from the ecosystem. The launch is part of a broader infrastructure pivot led by Chief AI Officer Alexandr Wang, backed by Meta’s projected $115 billion to $135 billion capital expenditure for 2026. As you update your AI agent frameworks, you must factor in the transition to proprietary API endpoints for Meta’s latest capabilities.
If your production stack depends on open-weight Llama deployments, Muse Spark requires a strategic reassessment of your roadmap. The shift to a closed API model means you will need to rely on existing open models for local inference or migrate your workloads to Meta’s managed endpoints. When evaluating the new Contemplating Mode, measure its parallel orchestration against your current custom agent loops to determine if native model orchestration reduces your latency and complexity.
Get Insanely Good at AI
The book for developers who want to understand how AI actually works. LLMs, prompt engineering, RAG, AI agents, and production systems.
Keep Reading
Train Multimodal Sentence Transformers for Visual Retrieval
Learn how to finetune multimodal embedding and reranker models for text, image, and audio using the updated Sentence Transformers library.
Gemini 3.7 Flash Reaches Google Search AI Mode
Google has integrated Gemini 3.7 Flash into Search AI Mode for paid subscribers, adding configurable reasoning and a 1M-token context window.
Google Ships Gemini 3.7 Flash With Adjustable Reasoning Tiers
Google DeepMind's Gemini 3.7 Flash introduces tunable thinking levels and hits 65.3% on DeepSWE, targeting complex agentic workflows.
VoxParadox Benchmark Exposes Transcript Bias in Audio LLMs
USC researchers introduced the VoxParadox benchmark to demonstrate how advanced Audio LLMs ignore paralinguistic cues in favor of literal transcripts.
Gemini Omni Flash Powers Conversational Edits in Google Vids
Google Workspace has updated Vids with Gemini Omni Flash for conversational video editing and introduced SynthID-watermarked personal AI avatars.