Meta announced the launch of two new artificial-intelligence systems, Muse Wonder and Muse Wonder Full, on Monday. The company describes the offerings as its most advanced models yet, designed to move toward what it terms “personal superintelligence.”

Model Overview

The two models introduce features called Active Memory and Active Focus. Active Memory enables the system to retain pertinent details throughout a conversation, while Active Focus is intended to help the AI maintain attention on demanding assignments over longer periods. Together, the capabilities aim to improve continuity and depth in user interactions.

Performance Benchmarks

An external evaluation platform, Artificial Analysis, assigned the standard Wonder model a score of 52 on its Intelligence Index, and the full-feature version a score of 56, both measured at the highest reasoning effort. Those numbers place Wonder Full above Claude Fable 5.1 and GPT-6 Astra, each recorded at 53, but still behind Claude Opus 5.5, which achieved 58. Independent verification from Abacus.AI’s chief executive, Bindu Reddy, questioned the claims, noting on social media that Muse Wonder does not match the performance of Fable or Astra and promising detailed benchmark data later.

Pricing and Access

Meta priced the base Wonder model at $4 per million input tokens, aligning it with Claude Opus 5.5 and undercutting GPT-6 Astra’s $10-per-million rate and Claude Fable 5.1’s $50-per-million fee. Access to Wonder Full is restricted to subscribers of Muse Code’s usage plans, which start at $50 per month for a “power usage” tier and rise to $200 per month for a “maximum usage” tier.

Industry Reaction

The introduction of Muse Wonder adds another contender to the crowded generative-AI market, where pricing and performance remain key differentiators. While Meta’s pricing strategy appears aggressive, the mixed reception from independent evaluators suggests that market acceptance may hinge on proven capability. Analysts will be watching how the subscription model for the higher-tier offering performs relative to the pay-as-you-go structures of competing services.

Why it matters

Meta’s entry with Muse Wonder signals the company’s intent to compete directly with established AI providers such as Anthropic’s Claude series and OpenAI’s GPT line. By bundling novel context-handling features with a competitive price point, Meta could attract developers and enterprises seeking cost-effective, high-capacity language models. The outcome may influence pricing dynamics across the sector and shape the development of next-generation AI tools.