San Francisco hosted OpenAI’s annual DevDay event on Tuesday, September 29, 2026, where Chief Executive Officer Sam Altman unveiled major upgrades to the company’s ecosystem. The announcements featured new always-on autonomous AI agents called Dots, the cost-effective GPT-6.1 Sol model, and a high-performance Ultrafast speed tier designed to accelerate developer workflows.
The conference drew developers and industry experts for hands-on workshops and technical deep dives. According to the event’s official website, attendees gathered to test new tools, collaborate, and share feedback directly with the engineering teams building OpenAI’s core technologies.
Introducing Dots: Always-On Autonomous Agents
Powered by the flagship GPT-6 Astra model, Dots operate continuously on dedicated cloud environments. Unlike traditional conversational assistants that remain idle waiting for user prompts, Dots run background tasks 24/7 and integrate with over 4,000 external applications across platforms like Slack, Microsoft Teams, and ChatGPT.
During the keynote demonstration, Altman emphasized the shift in productivity, stating, “We think you are really going to love this and not want to go back.” Users can start with a single primary assistant and scale up to coordinated teams of agents handling complex multi-step workflows, from monitoring software bugs to updating sales proposals.
Safety and Privacy Architecture
To mitigate risks associated with unsupervised background operations, OpenAI implemented strict security frameworks for Dots:
- Read-Only Proactive Research: When working unprompted, agents are restricted to read-only access to scan for tasks without modifying content or sending messages.
- Auto-Review Gateways: Actions affecting user accounts undergo an automated review step against custom rules and safety parameters.
- Activity Tracking: Users maintain complete visibility through a dedicated dashboard to audit background work and redirect agents as needed.
Expanding Model Offerings with GPT-6.1 Sol and Ultrafast Tier
In addition to personal agents, OpenAI expanded its underlying architecture options. The newly introduced GPT-6.1 Sol model provides a budget-friendly alternative to the high-end Astra flagship, undercutting heavier models while maintaining practical capability for everyday tasks.
Performance upgrades also arrived via the Ultrafast speed tier, rolled out across ChatGPT, Codex, and the core API. This tier delivers token generation speeds up to eight times faster than standard baselines. Additionally, the company launched a Pro 500 tier catering to heavy enterprise users requiring expanded usage limits and higher resource allowances.
Availability and Rollout
Dots are currently rolling out to ChatGPT Pro, Business Premium, and Enterprise subscribers in eligible markets, with setup managed through the desktop app or browser interface. As OpenAI continues scaling its enterprise and developer tools, the integration of autonomous cloud agents signals a definitive shift toward continuous, background-driven artificial intelligence collaboration.
