Overview
Meta announced Muse Spark, their first model release in a year, competing with top-tier AI models while introducing two distinct reasoning modes on meta.ai. The model offers "Instant" and "Thinking" modes with a future "Contemplating" mode promised.
Key Facts
- Competitive with Opus 4.6, Gemini 3.1 Pro, and GPT 5.4 on benchmarks - Meta is now playing in the top tier of AI models
- Offers Instant and Thinking modes with Contemplating mode coming - users can choose speed vs quality for different tasks
- Behind on Terminal-Bench 2.0 and coding workflows - still gaps in programming capabilities compared to rivals
- Hosted model with private API preview - no open weights unlike previous Llama releases
- Includes 16 different tools including web browsing and Meta content search - can search Instagram, Threads, and Facebook posts since 2025
- Renders SVG and HTML inline like Claude Artifacts - enables interactive visual outputs directly in chat
- Thinking mode produces better quality outputs than Instant mode - visible quality difference between reasoning approaches
Why It Matters
This represents Meta's shift toward hosted premium AI services rather than just open-source releases, directly competing with OpenAI and Anthropic while leveraging unique access to Meta's social media data.