Author: Dong Jing
As Meta’s personal AI agent Muse receives JPMorgan’s highest rating and gains widespread attention, OpenAI is accelerating the launch of a persistent AI assistant named "o" to directly compete in the increasingly fierce AI agent race.
According to tech media WccfTech, OpenAI has added the "o" logo to the product documentation for ChatGPT Pro, indicating that the official launch is imminent and expected to take place at the Dev Day event this Tuesday (September 29).
This move is widely interpreted as a direct response to Meta Muse’s recent strong performance—J.P. Morgan ranked Meta Muse as the top AI agent in its latest Assistant Benchmark evaluation.
This Dev Day will also bring more major announcements: OpenAI is expected to simultaneously launch a new subscription tier, ChatGPT Pro Max, priced as high as $500 per month, offering ultra-fast response experiences. Meanwhile, "o" is rumored to work in tandem with a consumer-grade AI hardware device from OpenAI, shaped like a donut and featuring moving parts, further expanding its product ecosystem.
"o": Always online, designed for long-running tasks
Reports state that in response to the challenge from Muse, OpenAI has launched a product called "o," positioned as a "persistent AI assistant."
Public information is currently limited, but @synthwavedd posted on social platform X that "o" is reportedly powered by a specific variant of GPT-6 Astra called "Aeon," which is said to excel in handling long-running tasks—the very core capability required for an efficient personal agent.
In the competitive landscape, "o" faces not only Muse but also OpenAI, which has entered the arena targeting both Grok Bot and Hermes Agent. With three players vying for dominance, competition in the AI personal assistant sector is rapidly intensifying.
Dev Day Strikes Dual Fronts: New Brand + Premium Subscription
Besides "o," OpenAI is expected to roll out several major updates simultaneously at Dev Day, creating a powerful combination.
At the subscription pricing level, OpenAI plans to launch ChatGPT Pro Max with a monthly fee of up to $500, offering ultra-fast response times—reportedly powered by Cerebras’ Wafer-Scale Engine—targeting professional users with extremely high demands for speed and performance.
On the hardware ecosystem level, "o" will also serve as a software companion, deeply integrated with OpenAI’s upcoming consumer-grade AI device. The device is rumored to resemble a donut and may feature moving parts, signaling OpenAI’s continued ambition in the AI hardware space.
Reports indicate that both product lines are being advanced simultaneously, suggesting that OpenAI is attempting to build competitive barriers on two fronts—software agents and hardware endpoints—to counter multifaceted pressure from competitors like Meta.
JPMorgan crowns Muse, putting immense pressure on OpenAI
The rise of Meta Muse was the direct catalyst for OpenAI's urgent response.
According to the latest Assistant Benchmark evaluation by J.P. Morgan, Meta Muse ranked first among all evaluated AI agents, outperforming multiple competitors, including OpenAI’s products. This authoritative endorsement has continued to draw market attention to Muse and has, to some extent, captured the pre-launch buzz that was originally intended for OpenAI Dev Day.
Muse is a personal AI agent designed to perform multi-step tasks on behalf of users, with core capabilities including autonomously launching a cloud-based browser, browsing the web like a human, and continuing to process tasks in the background even after the app is closed.
To address privacy concerns, Muse runs on a dedicated, isolated cloud virtual machine (VM) with a separate browser for securely storing user data and login credentials.