What happened
Meta introduced Muse Wonder, codenamed Watermelon, as the latest generation of its Muse personal‑assistant AI. The company says the model improves task completion for “personal and professional agents,” handling real‑world obstacles such as restaurant reservations and availability checks without stalling. Access to Muse Wonder is offered through the Muse app via two paid tiers – a $20 per month plan and a premium $100 per month plan – both of which unlock the model’s full capabilities and its “strongest safeguards to date” against cyber misuse. A more powerful version, Muse Wonder Full, is available only to developers who subscribe to Meta’s Muse Code platform for $50 or $200 per month. The model will also be accessible through Meta’s Model API, though at a higher price point than the consumer plans. Meta AI chief Alexandr Wang highlighted that the company is monitoring “agent‑related incidents” after some users reported the agent accessing private messages, a claim Meta disputes. The launch follows Meta’s broader push to generate AI‑derived revenue beyond its core advertising business.
Meta announced the Muse Wonder model at its Meta Connect event on September 23, 2026, describing it as the most advanced option for powering digital assistants. The model, internally known as Watermelon, was trained to handle real‑world task interruptions and to propose alternative approaches without stalling.
Consumers can unlock Muse Wonder through the Muse personal‑assistant app by subscribing to either a $20 or $100 monthly plan. Both plans grant access to the model’s full capabilities and its latest safety mechanisms, which Meta says are the strongest it has deployed to date.
Developers and power users have a separate offering: Muse Wonder Full, available via the Muse Code platform for $50 or $200 per month. This variant is described as a “much larger model at frontier level,” and will also be reachable through Meta’s Model API at a higher price point.
Meta AI chief Alexandr Wang noted that the company is actively monitoring agent‑related incidents. Some users have alleged that Muse accessed private messages without permission; Meta has disputed those claims and is working on user education to mitigate such concerns.
The launch is framed as part of Meta’s broader effort to generate AI‑derived revenue streams beyond its core advertising business, positioning the subscription model as a competitive alternative to offerings from OpenAI and Anthropic.
Why it matters
The Muse Wonder rollout marks Meta’s first monetized AI model for consumer assistants, signaling a strategic shift toward subscription revenue as the company seeks to diversify away from ad‑driven income. By pricing the model below competing offerings from OpenAI and Anthropic, Meta aims to attract price‑sensitive users while positioning itself as a viable alternative in the personal‑assistant market. The introduction of stronger safeguards, albeit only for paying users, raises questions about the balance between safety and accessibility, especially after prior complaints of the agent reading private messages. If the subscription model gains traction, it could influence how other large tech firms package AI capabilities, potentially accelerating a broader industry move toward paid AI services. Moreover, the developer‑focused Muse Wonder Full tier could spur new integrations and third‑party applications, expanding Meta’s ecosystem beyond its social platforms.
Revenue diversification: By monetizing Muse Wonder through subscriptions, Meta is testing a new business model that could reduce reliance on ad revenue, a strategic priority highlighted in recent earnings calls.
Pricing competition: Meta claims its subscription rates are “meaningfully below” those of leading rivals, potentially reshaping price expectations for consumer‑grade AI assistants.
Safety access: The decision to tie the strongest safeguards to paid tiers raises policy questions about equitable access to protective features, especially for users who cannot afford the premium plans.
Developer ecosystem: The Muse Wonder Full tier could catalyze third‑party development, expanding Meta’s AI footprint into new applications and services beyond its social platforms.
User trust: Ongoing complaints about unauthorized message access, even if disputed, underscore the importance of transparent privacy practices for AI assistants.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
crm_get_transaction(id='4092').Which component of an AI application is the machine-learning model itself?
What to watch next
Key indicators to monitor include subscription uptake rates, especially the premium $100 tier, and user sentiment regarding privacy and the reported message‑access issue. Analysts will watch whether Meta’s pricing truly undercuts OpenAI and Anthropic in practice, factoring in any hidden costs for API access. The performance and adoption of Muse Wonder Full among developers could reveal the model’s appeal for enterprise‑level tasks. Regulatory scrutiny may also intensify if safety features remain gated behind a paywall, prompting potential policy discussions about equitable access to AI safeguards. Finally, Meta’s ability to translate Muse Wonder usage into measurable revenue will be a barometer for the viability of AI‑centric subscription models across the tech sector.
Subscription uptake: Tracking the number of users opting for the $20 versus $100 plans will indicate market appetite for paid AI assistants.
Developer adoption: Monitoring API usage and the number of applications built on Muse Wonder Full will reveal the model’s appeal to the developer community.
Privacy incidents: Any further reports of unauthorized data access could trigger regulatory attention or impact user confidence.
Competitive response: Reactions from OpenAI, Anthropic, and other AI providers to Meta’s pricing strategy may lead to price adjustments or new feature rollouts.
Regulatory developments: Policymakers may scrutinize the practice of gating safety features behind a paywall, potentially prompting new guidelines for accessibility.