
Tech • AI • Robotics • Game
OpenAI unveiled a new personal assistant product called DOTS, a collaborative workspace named Space, a lower-cost model called GPT-6.1 Soul, a faster Decisions API, and a new $500-per-month plan with sharply higher usage limits.
DOTS is designed as a personal AI assistant that can be contacted across multiple channels, including in-app chat, phone calls, SMS, email and Slack. Unlike a standard chatbot that waits for prompts, it can proactively surface updates, such as summarizing email and Slack activity and flagging urgent items before a user asks.
Users can create and name separate DOTS for different tasks rather than relying on a single assistant. The product is positioned as a set of specialist agents that can manage ongoing work, follow up independently and remain available continuously instead of only responding inside a single chat window.
The assistant combines Codex, ChatGPT Work, a cloud browser and connected third-party apps. That setup allows the agent to take actions in a cloud-based environment rather than on a user’s local machine, letting it browse sites, review tools and complete tasks while the user continues working elsewhere.
OpenAI is allowing users to restrict which apps a DOT can access and to require approval before immediate actions are carried out. Those controls are intended to limit unintended behavior while still enabling practical automation across email, messaging and other connected software.
A first DOT is included at no extra cost for Pro or Business Premium subscribers. Regular conversations with a DOT do not count against standard ChatGPT usage limits, but actions that invoke Codex or ChatGPT Work do consume usage credits.
OpenAI also launched Space, a new collaborative workspace that resembles a unified document and project environment for teams and assistants. It is designed so users and DOTS can work on the same pages, with assistants able to create dashboards, update documents and respond to tagged requests inside a shared interface.
The new GPT-6.1 Soul model is being pitched as close to GPT-6 Astra in intelligence while costing far less. Pricing presented for developers shows $2 per million input tokens versus $10 for Astra, $1 cached input versus $10, and $10 output versus $50, making Soul roughly one-fifth the price.
GPT-6.1 Soul is available across ChatGPT Work, Codex and the API for Plus, Pro, Business and Enterprise users. The release comes amid an increasingly aggressive competition in frontier AI, where major vendors are shipping new flagship and mid-tier models in rapid succession.
OpenAI also introduced a $500 monthly subscription tier that includes the new Ultrafast model, removes the five-hour limit and offers 25 times the usage of the Plus plan. The plan also extends usage across partner applications, signaling a push toward more intensive professional and enterprise workflows.
The new Decisions API is intended to help AI agents choose actions faster, improving performance in computer-use tasks such as navigating interfaces and clicking through booking flows. Faster decision-making could address one of the main weaknesses of earlier agent systems, which often felt too slow for practical day-to-day automation.
The announcements show OpenAI pushing beyond chat into always-on assistants, shared workspaces and higher-volume professional usage. The central test will be whether DOTS, Space and the new pricing tiers deliver reliable automation without adding complexity or risk.
Ask a question