Meta claims new Muse Wonder AI models rival Open
Reported by Cointelegraph · AI-assisted summary by ChikoCorp AI News Desk
AI-generated summary based on the linked source; not independently verified. This is not investment advice. Verify market-moving details at the original publisher before acting. See our editorial policy, AI content policy, and financial disclaimer.
Summary
Meta introduced two new AI models, Muse Wonder and Muse Wonder Full, which it describes as its most advanced to date and a step toward “personal superintelligence.” These models feature capabilities called Active Memory and Active Focus to enhance conversation continuity and task persistence. According to an independent evaluation by Artificial Analysis, Muse Wonder Full scored 56 on its Intelligence Index, outperforming some competitors but trailing Claude Opus 5.5, while Meta’s performance claims were disputed by Abacus.AI's CEO.
Why it matters
The source implies Meta's development of advanced AI models could influence the competitive landscape of AI technologies, particularly in high-performance, personalized AI applications. However, the article notes disagreement over the models' performance levels, indicating ongoing debate within the industry. The source does not directly elaborate on broader market or policy impacts.
Key context
Meta highlights new features Active Memory and Active Focus as core innovations in the Muse Wonder models, enabling information retention across conversations and sustained focus on complex tasks. The pricing positions Meta’s standard model competitively below some rivals, and the full model is restricted to higher-tier subscribers of Muse Code. The source mentions independent evaluations and critiques from industry experts but provides limited technical detail.
Key numbers and entities
Meta, Muse Wonder, Muse Wonder Full, Artificial Analysis Intelligence Index scores (Wonder Full 56, Wonder 52), Claude Fable 5.1 (53), GPT-6 Astra (53), Claude Opus 5.5 (58), Abacus.AI CEO Bindu Reddy, Muse Code subscription tiers ($50 and $200 per month), pricing per million tokens ($4 for Wonder, $10 for Astra, $50 for Fable).
What remains unclear
The source does not clarify the specific methodologies behind the Intelligence Index scores or the full spectrum of capabilities distinguishing Muse Wonder from competitors. It is also unclear how broadly available the models are beyond Muse Code subscribers or how Meta intends to deploy these models commercially. Detailed benchmark results and independent analyses are pending.