Meta AI announced
Meta introduced an assistant across its messaging and social apps, initially powered by Llama-family technology.
- Group-chat access
- Web information
- Image creation
- Multiple Meta app surfaces
A personal assistant embedded across Meta apps, the standalone app, web and AI glasses, with social context and creative media tools.
Meta AI is the consumer assistant. Muse Spark 1.1 powers current reasoning and agent features; Muse Image handles generation/editing, while future media and open-model releases follow separate tracks.
Meta AI launched inside WhatsApp, Messenger and Instagram, then expanded to Facebook, the web, a standalone app and Ray-Ban/Oakley devices.
The current Muse stack shifts the assistant from answering and generating images toward planning, app connections, recurring tasks, research and deliverable creation.
Product features and underlying models are documented separately. Roadmap items include only provider-announced work; rumors are excluded.
Selected milestones that materially changed the product.
Meta introduced an assistant across its messaging and social apps, initially powered by Llama-family technology.
Meta expanded country availability, added meta.ai and made search/feed access and image generation faster.
A rebuilt multimodal model added tool use, visual reasoning, parallel subagents and faster live multimodal interaction.
Meta AI added planning, connected apps, recurring work, deep research, slide creation and an agentic image model.
The comparison uses provider-published capabilities and named evaluations, not a generic marketing score.
| Dimension | Previous | Current | Measured difference |
|---|---|---|---|
| Primary behavior | Answers, web results and creative generation | Plans, connects to apps and follows through on tasks | Assistant becomes more agentic |
| Reasoning | Single-assistant Llama 3 workflow | Instant/Thinking modes with parallel subagents | Complex prompts can be divided into concurrent workstreams |
| Multimodality | Image prompts and generation | Live camera understanding plus Muse Image editing and composition | From static generation to perception and iterative creation |
| Deliverables | Chat answers and images | Research reports, slides, plans and mood boards | More work products persist in one place |
Official launch material describes capability and product changes but does not provide enough directly comparable scores to claim that Muse Spark 1.1 is a specific number of times stronger than the prior Meta AI.
Muse Spark can launch multiple subagents for different parts of a request and combine the result.
Check evidenceMeta says the assistant is being distributed across apps used by billions; reach is not a capability benchmark.
Check evidenceNo official apples-to-apples product score supports a single multiplier.
What happens behind the button, where it helps and what to verify.
How it works: The assistant creates a plan, uses connected context, continues steps and can deliver recurring briefings or updates on a schedule.
Best use: Training plans, event planning and daily briefings
Watch: Review connected-account permissions and any suggested action.
How it works: Retrieves public posts, Reels, creator content, maps, Marketplace and web options to enrich recommendations.
Best use: Local discovery, trends, styling and product research
Watch: Recommendations may reflect popularity or available public data rather than objective quality.
How it works: Processes live camera and voice context so the assistant can respond to what the user is seeing and hearing.
Best use: Hands-free assistance and visual questions
How it works: Plans a generation, can invoke search or code, blends references and supports conversational or drawn-on-image edits.
Best use: Visual concepts, edits, infographics and social media
Watch: Generated text, objects and public-context references should be checked before publication.
Status reflects the provider's latest public wording. “No public date” means there is no verified launch schedule.
Planning, connected-app and action features are moving to more countries and surfaces, including WhatsApp.
Official announcementMeta says Muse Video is in development and coming to creators and Meta AI; no fixed public release date was given.
Official announcementMeta says larger models are in development and hopes to open-source future versions, without a guaranteed schedule.
Official announcementProvider announcements, help documentation and system information.
Meta · Apr 18, 2024
Meta · Apr 8, 2026
Meta · Jul 24, 2026
Last editorial verification: 2026-08-24. Prices, access limits, preview status and model routing can change after this date.