Meta's frontier model for personal and professional AI agents, with Active Memory and Active Focus; successor to Muse Spark.
Updated EU status checked Last shown
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
Muse Wonder is a frontier model from Meta, presented on Meta AI Research as a model for personal and professional AI agents. Meta describes it as keeping track of what matters to the user through "Active Memory" and "Active Focus", so it can work toward goals at work and in daily life. It follows the Muse Spark line within Meta's Muse model family, which also includes Muse Ember, Muse Glimmer, Muse Code, Muse Image and Muse Video.
A YouTube creator summary describes it as fast and less prone to hallucination than earlier models, but expensive per task, and ranking around fifth place or lower on several independent benchmarks. These are secondhand impressions and not figures from Meta's pages.
The scraped page gives no details on pricing, API access, context length, license or regional availability. Exact release date, modalities and access channels are not confirmed in the material provided.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Meta's new frontier model for agentic and professional work, succeeding Muse Spark. It is fast and hallucinates less, but is expensive per task and ranks around top five or lower on several independent benchmarks.
Meta's new state-of-the-art model at $4/$20 per million tokens, scoring 79% on a coding benchmark and near top-tier on Artificial Analysis. The creator's tests showed a working Mega Bonk game with weaker graphics, and ranked 14th on his Beauc leaderboard.
Meta's new frontier model released the day before. The creator says it is strong at coding and tops Deep Sweep at 79%, but is weak at 3D. Alexander Wang said the team did not RL it for 3D.
Meta's frontier model, available via API at $4/M input, $20/M output, $0.20/M cached input. Benchmarks place it near the top (second-tier on Artificial Analysis index at 56), and the creator found it strong on benchmarks but expensive per task.
Closest neighbours by meaning, not just by category.
Model · Meta
Closed-source Meta model for coding and agentic workflows with 1M-token context and multimodal input, offered via API.
Model · Meta
Meta's Muse Spark 1.3 language model, described by creators as near-frontier on benchmarks.
App · Meta
Meta's personal AI assistant app with voice conversations, expressive avatars and task handling across connected apps.
Model · Meta
Meta's smallest Muse-series language model, positioned as a low-cost option for subtasks and delegation.
Model · Meta
Meta's proprietary large language model, cited by creators as a top US-made model on the Artificial Analysis index.
Model · Meta
Meta's open-weights 30B dense model for local agents, tuned for tool use and long tasks, Apache 2.0 licensed.
