OpenAI’s DevDay Puts Always-On Dots Beside Cheaper GPT-6.1 Sol

OpenAI described DevDay 2026 as its biggest yet, with more than 20 major announcements spanning ChatGPT, Codex, models, and new ways of working with AI. [1] At the event, the company introduced agents designed to take on ongoing responsibilities and new ways for people and AI to work together. [1]

OpenAI said ChatGPT has 1.2 billion weekly users and is being opened as a shared surface where people and agents can collaborate and developers can launch native experiences. [1] OpenAI also launched Dots, a personal agentic assistant powered by GPT-6 Astra and designed to pursue user-defined goals continuously in the background with minimal oversight. [3]

OpenAI says users can begin with a primary Dot, give it a name, and personalize it, while the company envisions teams of Dots eventually working together for users. [3] The company’s examples include a developer assigning a Dot to monitor customer feedback and implement fixes, and a scientist using one to rerun analyses when new experimental data arrives. [3]

Users can message Dots through Slack, Teams, and other organizational platforms, with text-message support planned for later. [3] Individual Dots can receive specific identities, credentials, and tools through existing systems, and OpenAI is working with Microsoft on integration with Agent 365 security controls. [3] The Verge positioned Dots as a rival to Meta’s recently launched Muse and noted that Muse is free while Dots initially targets paid ChatGPT plans. [6]

OpenAI presented GPT-6.1 Sol at DevDay one week after launching GPT-6 Sol. [4] The company says GPT-6.1 Sol delivers nearly the same intelligence as GPT-6 Astra for agentic coding, computer use, and professional work at one-fifth of Astra’s standard input and output token prices. [4]

At low reasoning effort, OpenAI says the share of responses containing a factual error fell from 11.4 percent with GPT-6 Sol to 7.7 percent with GPT-6.1 Sol. [4] Across all reasoning settings, the company says GPT-6.1 Sol’s error rate remains within 1.9 percent of GPT-6 Astra, and the model is available in ChatGPT Work and Codex but not yet in Chat. [4]

OpenAI introduced reusable Codex cloud development environments that developers can access from different devices, extending the engineering agent beyond a developer’s laptop. [2] While Codex already supported remote cloud tasks, the new environments are more persistent and configurable instead of treating each task as an isolated remote sandbox. [2]

OpenAI says the environments should let tasks start faster and provide teams with shared workspaces containing approved settings and permissions. [2] The refreshed Codex CLI adds voice-directed tasks and an /agents view for delegating and tracking concurrent work, while the ChatGPT desktop app gains code-review summaries, change exploration, issue questions, and automatic initial reviews. [2]

Dots packages persistent, independently acting assistance into a named product, while Muse supplies the clearest competitive reference: Meta’s product is free, whereas Dots initially sits behind paid ChatGPT plans. GPT-6.1 Sol offers a different trade-off, aiming to approach GPT-6 Astra across agentic work while charging one-fifth of its standard token prices. Reusable Codex environments complement that model strategy by giving coding agents persistent configurations and shared permissions across devices. Together, the announcements make continuity—of goals, environments, credentials, and work—the defining feature of OpenAI’s agent push. What the sources don't address: how OpenAI will measure and constrain failures when Dots and Codex act continuously with organizational credentials and limited oversight.