Meta's Muse AI Agent Outpaces ChatGPT's Early Growth as OpenAI and Anthropic Trade Model Drops

Meta's personal AI agent Muse is pulling in users at a faster clip than ChatGPT did in its first months, according to early figures reported by TechCrunch, and the company is preparing to push the assistant onto smart glasses. The traction landed on the same day Anthropic released Opus 5.5 and OpenAI followed with GPT-6 model updates roughly 90 minutes later, a back-to-back cadence that turned the week into an unofficial model drop window.

The timing is awkward for the two labs that spent much of the past year publicly discussing 'pacing the frontier,' a phrase their executives have used to describe a deliberate, safety-conscious release rhythm. Whatever the internal schedule looks like, the public result was a compressed race: Anthropic's Opus 5.5 arrived first, OpenAI's GPT-6 refresh followed almost immediately, and neither generated the chatter that Meta's consumer-facing agent did. That contrast matters because frontier-lab prestige has rested on benchmark leadership, while Meta is competing on daily habit instead.

Muse is a different kind of product from a chatbot sitting in a browser tab. It is built as a persistent personal agent, the sort of companion-style assistant that lives on a phone or a wearable and accumulates context about its owner over time. Meta's plan to extend it to smart glasses puts the agent directly into the Ray-Ban line the company has been shipping for years, a hardware channel OpenAI and Anthropic do not control. If the early numbers hold, Meta would own the distribution layer for consumer AI rather than renting it through someone else's app store.

The comparison to ChatGPT's early growth is the number that will get quoted, but it deserves scrutiny. ChatGPT launched into a market with almost no direct competition and rode a wave of novelty; Muse arrives into a crowded field where every phone ships with an assistant and every lab has a consumer app. A faster adoption curve in 2026 is a different achievement than the one OpenAI posted in late 2022, and Meta has not disclosed retention, session length, or how many Muse users are simply tapping a preinstalled icon rather than seeking the product out.

What Meta does have is reach. Its apps already touch billions of accounts, and a Tamagotchi-style agent that users check on daily fits the engagement patterns the company has spent two decades optimizing. That machine can make an average product look huge and a good product look unstoppable, which is why the Muse numbers are being read as a strategic threat rather than a curiosity.

The next signals to watch are concrete. Meta needs to show retention for Muse past the first month and a working glasses integration, not just a demo. Anthropic and OpenAI, meanwhile, have to decide whether rapid-fire releases answer a distribution problem or just confirm that the frontier race is now measured in consumer downloads. The labs that talked about pacing have not explained what happened to the pace.