OpenAI Accelerates ‘o’ AI Agent Project to Counter Meta Muse

Ender Öztürk

September 28, 2026

Following the significant market success of Meta’s Muse personal AI agent, OpenAI is rapidly accelerating the development of its own competitor, internally referred to as the “o” AI agent. The company is expected to officially unveil this new assistant during its upcoming Dev Day event scheduled for next week. By introducing “o,” OpenAI intends to fortify its competitive position in the rapidly evolving artificial intelligence landscape while simultaneously diversifying its subscription ecosystem. This strategic move signals a direct challenge to existing industry players, as OpenAI aims to integrate this always-active assistant into a broader range of hardware and software solutions designed to enhance user productivity.

“o” and ChatGPT Pro Max Plans Are Evolving

The “o” project represents a critical pillar in OpenAI’s strategy to maintain its dominance. Reports suggest that the agent is powered by a variant of the Astra model, known as “aeon,” which has been optimized for handling persistent, long-duration tasks. Industry analysts believe this software will serve as the primary companion for OpenAI’s upcoming consumer hardware, rumored to feature a unique, donut-shaped design. Development on this technology has been underway for several months, suggesting a coordinated push to reclaim market share.

Alongside the agent, the company is preparing to launch the ChatGPT Pro Max plan. Priced at $500 monthly, this premium service is set to utilize high-performance Wafer-Scale Engine infrastructure provided by Cerebras. This tier promises significantly faster response times and increased usage quotas compared to standard offerings. By segmenting its services, OpenAI hopes to cater to power users who require reliable, low-latency performance for heavy workloads.

Meta Muse Sets New Standards for Personal Assistants

The urgency behind OpenAI’s developments stems from the impressive performance of Meta’s Muse. According to recent evaluations by JP Morgan, Muse achieved an industry-leading score of 9.3 on the Assistant Benchmark. The system demonstrated exceptional proficiency in complex domains, including online purchasing, email management, and automated routine execution. Notably, Muse operates within an isolated virtual machine, utilizing a cloud-based browser to navigate websites autonomously, a capability that marks a significant leap in agentic technology.

The speed advantage demonstrated by Meta’s solution is substantial. While conventional assistants often require nearly 38 seconds to process multi-step requests, Muse completes similar tasks in approximately 7 seconds. This efficiency gap has positioned Meta as a formidable force in the personal assistant sector, forcing competitors to rethink their development timelines. As the industry shifts from simple chatbots to proactive agents, the ability to interact with external platforms securely and rapidly has become the primary metric of success.

OpenAI’s upcoming hardware and software integration aims to bridge this gap, ensuring that users can manage their daily routines with greater efficiency. With the introduction of the “o” agent and the high-performance Pro Max tier, the stage is set for a major showdown in the AI market.

Given the rapid advancements in agentic AI and the upcoming release of OpenAI’s “o” project, do you believe these new tools will successfully surpass the efficiency currently offered by Meta Muse? Share your thoughts on the future of personal AI assistants in the comments section below.