OpenAI Launches dots: Always-On GPT-6 Astra Agents Working in Cloud
OpenAI rolls out dots, always-on AI agents running on dedicated cloud computers for Pro and Business users, powered by GPT-6 Astra achieving 72.6% on OSWorld 2.0.
Stock photo for illustration only, not from the actual event
- OpenAI launches dots, persistent AI agents running on dedicated cloud computers
- Powered by GPT-6 Astra, scoring 72.6% on OSWorld 2.0 benchmarks
- Available now for ChatGPT Pro and Business Premium users in eligible markets
- Each user starts with 1 primary dot with a new $500 tier introduced
OpenAI has officially launched dots, an always-on AI agent system operating on dedicated cloud computers. This transition shifts ChatGPT from a simple request-and-response loop into a standing assignment tool where users can assign a goal, name, and boundaries. The system then pursues objectives around the clock while learning from user feedback, with each user starting with one primary dot.
Regarding deployment, dots are rolling out as a managed product for ChatGPT Pro and Business Premium users in eligible markets. Enterprise, Edu, and Healthcare workspaces gain access through a beta once administrators enable it, with no self-hosted or open-weights options available. The first dot is included at no extra cost in Pro and Business Premium plans, and conversations do not count toward general ChatGPT usage limits unless tasks are initiated in Codex or ChatGPT Work.
Stock photo for illustration only, not from the actual event
Under the hood, the system is powered by advanced models. In OpenAI's latency simulations, GPT-6 Astra scores 72.6% on OSWorld 2.0 taking roughly 40 minutes per task, compared to its predecessor GPT-5.6 Sol which scored 65.7% at about 75 minutes per task. OpenAI also rolled out a $500 paid tier on the exact same day for advanced users.
OpenAI also released details in a separate safety post, noting that dots can still make mistakes and consequential work requires human review. This caution is timely; a day before launch, OpenAI revealed that its agents inadvertently posted users' images online, affecting 53 ChatGPT users.
The introduction of dots represents a major shift toward autonomous agent ecosystems where AI actively performs multi-step workflows rather than just answering prompts. By allocating dedicated cloud infrastructure per user, OpenAI addresses environment isolation, but safety guardrails and accuracy remain paramount as these systems gain autonomous access to enterprise tools and data.
Furthermore, OpenAI is previewing specialist dots equipped with individual identities, credentials, and company system access for fixed roles. Having tested them internally across procurement, invoice processing, email marketing, customer support, and commercial contracting, the rollout begins with enterprise pilots alongside Microsoft integrations for Agent 365.
Source: MarkTechPost
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment