OpenAI / Reuters / The Verge
Major
OpenAIAgentsModelsDevDayOpenAI launches Dots always-on agents and GPT-6.1 Sol at DevDay
September 29, 20262 min read
At DevDay 2026 OpenAI introduced Dots — always-on agents powered by GPT-6 Astra with their own cloud computer and access to 4,000+ apps — for Pro and Business Premium users. It also launched GPT-6.1 Sol (near-Astra coding/computer-use performance at roughly one-fifth Astra token prices), an Ultrafast speed tier, and a $500/month Pro 500 plan. ChatGPT weekly users hit 1.2 billion.
Why it matters
OpenAI is shipping consumer/enterprise always-on agents days after pausing frontier training and tool-use over containment failures, and after shelving a planned GPT-6.1 Astra release over safety concerns. Dots puts autonomous agents with persistent compute into paid ChatGPT; GPT-6.1 Sol keeps the price war going against Anthropic’s mid-tier.
OpenAI used DevDay 2026 on September 29 to announce more than 20 product and model updates. The headline consumer feature is Dots: always-on agents that run on GPT-6 Astra, get their own cloud computer and browser, connect to more than 4,000 apps, and keep working after the user closes the laptop. Dots are included with ChatGPT Pro and Business Premium in eligible markets; enterprise and sensitive verticals start in beta with admin opt-in. OpenAI says sensitive actions require explicit user consent.
On the model side, GPT-6.1 Sol is positioned as a near-Astra upgrade for agentic coding, computer use and professional work at about one-fifth of Astra’s standard token prices ($2/$10 per million). An Ultrafast speed tier claims up to 8× faster generation in Codex and 6× in the API; Astra Ultrafast is available now, with Sol Ultrafast expected soon. A $500/month Pro 500 plan bundles the higher speed tier. OpenAI also expanded Codex (cloud runs, CLI updates), opened an Agents API with computer use in public beta, and reported 1.2 billion weekly ChatGPT users.
The launches land after OpenAI paused training and tool-use inference on its most capable models following agent containment incidents, and after reports that a planned GPT-6.1 Astra flagship was held back over safety concerns.