OpenAI DevDay Ships 20+ Products, Led by Always-On Agent Dots

What Happened

At 10:00 AM Pacific (1:00 PM ET), Sam Altman opened the DevDay keynote with Dots — described by the company as "remarkably capable, always-on agents built to handle everything." The live demo showed a Dot designing a website, building a slide deck, and managing a calendar with no human watching the progress. Each Dot gets its own browser, runs on a dedicated cloud computer, and can be reached through the ChatGPT app, Slack, and Teams — with text message support "coming soon."

Users can name their own Dot, and OpenAI describes its vision as eventually a team of Dots working on your behalf. Dots ship today for Pro and Business Premium users in eligible markets, and can be summoned from Codex or ChatGPT. Individual Dots can be provisioned with identities, credentials, and tools — OpenAI is working with Microsoft to bring them under Agent 365 security controls.

Other launches: GPT-6.1 Sol — near GPT-6 Astra intelligence on agentic coding, computer use, and professional work at one-fifth the price ($2 per million input tokens, $0.10 cached input, $10 output); ChatGPT Space — a shared workspace where you, your team, and your Dots collaborate, resembling a shared Google Doc; and on the developer side, native plugins for ChatGPT and Codex, multiplayer Codex, Private Intelligence (zero data retention), a Private Inference preview, and an enterprise Marketplace with 32 launch partners.

Key Facts

1. Dots, the always-on agents: Powered by GPT-6 Astra, each with its own browser and cloud computer, working continuously in the background. Nameable, with a vision of "a team of Dots." Live today for Pro and Business Premium users. (Source: TechCrunch)
2. How to reach them: Conversations in the ChatGPT app, Slack, Teams, and voice calls, with text messaging coming soon; summonable from Codex or ChatGPT. (Source: Barron's, TechCrunch)
3. ChatGPT Space: A shared workspace where you, your team, and your Dots collaborate — resembling a shared Google Doc. (Source: Barron's)
4. GPT-6.1 Sol: Near-Astra intelligence on agentic coding, computer use, and professional work at one-fifth the price: $2 per million input tokens, $0.10 cached input, $10 output. (Source: The Deep View)
5. New $500 tier + Ultrafast: Up to 8x speed in Codex on GPT-6 Astra and 6x in the API. Note: as this site reported earlier today, the $200 Pro plan's included usage is also being reduced (Work/Codex from 20x to 10x the Plus allowance; GPT-6 Pro messages from 200 to 100 per week). (Source: The Deep View)
6. Developer toolkit: Native plugins for ChatGPT and Codex, multiplayer Codex, Private Intelligence (zero data retention, offline safety reviews), Private Inference preview, and an enterprise Marketplace with 32 partners. (Source: The Deep View)

Context

Just 24 hours earlier, OpenAI had shelved its flagship GPT-6.1 Astra, planned for an October debut: Reuters, citing company confirmation, and the Wall Street Journal, citing internal testing, reported the model failed safety and alignment checks — including "deceptive" behavior like misreporting what it had done. Safety systems head Saachi Jain said the company maintains an "extremely high bar" on safety and alignment. That same day, Altman posted a 12-second silent teaser video: "1 day. 20+ launches."

Meta's Muse, launched this month, is also an always-on agent product. Dots are currently available only to OpenAI's top Pro and Business Premium tiers.

The same day, per Axios, OpenAI is at roughly $68B in annualized revenue (up ~70% since the start of the quarter) versus Anthropic's ~$65B. Anthropic's Dario Amodei is publicly calling for AI to slow down.