Product · October 3, 2026

Also published in Deutsch, עברית, Português (Brasil), العربية, Nederlands, Svenska, Norsk, Русский, Español, Italiano

OpenAI Unveils Always-On Dots Agents and GPT-6.1 Sol Model at DevDay 2026

Screens display coding text, representing programming work
Jakub Żerdzicki / Unsplash

OpenAI introduced a new suite of artificial intelligence agents named Dots and a cost-effective language model called GPT-6.1 Sol during its DevDay 2026 event. The company stated that these agents are designed to work continuously on user-defined goals, operating on cloud computers and continuing tasks in the background as needed. ABP Live Tech reported the story, detailing the specific capabilities and pricing structures of these new offerings. The Dots agents are built upon the GPT-6 Astra foundation and can integrate with more than 4,000 applications. They are capable of understanding user preferences based on feedback from these connected apps and can conduct research in the background, although currently, Dots can only read from connected applications during research phases.

Users can access Dots through ChatGPT, Slack, and Microsoft Teams. OpenAI emphasized user control, allowing individuals to set rules for which tasks the agents can perform autonomously. Sensitive actions, such as changing passwords, remain under direct user control and cannot be executed automatically by Dots. The company is also testing specialized enterprise agents for specific tasks like purchasing and invoice processing. OpenAI is collaborating with Microsoft to integrate these agents with enterprise security and management systems via Agent 365. Alongside the agents, OpenAI launched GPT-6.1 Sol, described as an upgrade to GPT-6 Sol. The company claims this model delivers performance close to GPT-6 Astra in agentic coding, computer use, and professional tasks at a significantly lower cost.

The price for GPT-6.1 Sol is up to one-fifth that of GPT-6 Astra. Specifically, the input cost for GPT-6.1 Sol is 2 dollars per 1 million tokens, and the output cost is 10 dollars per 1 million tokens. In comparison, GPT-6 Astra costs 10 dollars per 1 million tokens for input and 50 dollars per 1 million tokens for output. OpenAI also launched ChatGPT Space for teams, enabling shared file and information work between ChatGPT and Dots. Developers can now create plugin extensions, interactive panels, and file viewers within ChatGPT. Additionally, the Agents API now includes computer use capabilities. OpenAI announced @ChatGPT integrations for Slack and Teams, Sign in with ChatGPT for 16 partner tools, and an OpenAI Marketplace featuring 32 partners.