CL-SN55 941 —
GPT-61S 885 -0.6%
GPT-56T 861 —
GPT-6S 825 -1.4%
CL-OP55X 819 -1.4%
MUSE-SPK 793 -1.4%
QWEN-38X 779 -2%
GPT-6A 778 -0.5%
CL-OP5H 712 -1.7%
CL-OP5X 710 -1.7%
KIMI-K3X 700 -2.5%
CL-OP46H 695 -1.7%
CL-FAB5H 693 -1.7%
CL-OP55H 693 -1.7%
CL-OP47H 684 -1.7%
GEM-37FH 649 -1.8%
GEM-38FH 634 -3.4%
GPT-56S 574 -0.5%
CL-OP47 540 -0.6%
INKL 531 —
GEM-3P 494 —
CL-OP46 490 —
CL-OP48 483 —
GEM-31P 459 -5.4%
GPT-6L 436 —
CL-SN55 941 —
GPT-61S 885 -0.6%
GPT-56T 861 —
GPT-6S 825 -1.4%
CL-OP55X 819 -1.4%
MUSE-SPK 793 -1.4%
QWEN-38X 779 -2%
GPT-6A 778 -0.5%
CL-OP5H 712 -1.7%
CL-OP5X 710 -1.7%
KIMI-K3X 700 -2.5%
CL-OP46H 695 -1.7%
CL-FAB5H 693 -1.7%
CL-OP55H 693 -1.7%
CL-OP47H 684 -1.7%
GEM-37FH 649 -1.8%
GEM-38FH 634 -3.4%
GPT-56S 574 -0.5%
CL-OP47 540 -0.6%
INKL 531 —
GEM-3P 494 —
CL-OP46 490 —
CL-OP48 483 —
GEM-31P 459 -5.4%
GPT-6L 436 —
← Back to feed

OpenAI Dots Give Every User an Always-On Agent Backed by GPT-6 Astra

OpenAI used DevDay 2026 to announce Dots, an always-on agent layer inside ChatGPT available to Pro and Business Premium subscribers. Each dot runs on GPT-6 Astra, gets its own dedicated cloud computer and browser, and works toward assigned goals without waiting for the user to prompt it.

The product is OpenAI’s direct answer to Meta’s Muse, which launched in September 2026 as an ambient AI companion available on WhatsApp, iOS, Android, and the web. Where Muse targets consumer messaging, Dots spans ChatGPT, Slack, and Microsoft Teams from day one — targeting the enterprise workflows where OpenAI already counts 2.5 million business customers.

How Dots Work

Users assign a dot a multi-step goal and let it run. The dot has persistent memory — it learns preferences over time — and surfaces via voice call when human input is needed. Between check-ins, it handles the autonomous work: browsing, scheduling, research, execution through connected apps.

The cloud computer each dot gets is a meaningful commitment: OpenAI is provisioning compute per user rather than per query, a model more like a persistent worker than a chat session. Access comes through ChatGPT’s existing interfaces, with Slack and Teams integrations available at launch.

Sam Altman opened the DevDay keynote noting 1.2 billion weekly ChatGPT users and 2.5 million businesses on OpenAI, framing Dots as the next layer on top of that installed base.

Developer Surface

Alongside Dots, OpenAI released an Agents API with Computer Use, letting developers build agents that can interact with desktop applications directly. The Decisions API — a new routing primitive — handles intent classification and routes user requests to predefined actions, reducing the overhead of building structured agent pipelines from scratch.

Codex now runs in the cloud with a standalone interface, separate from the ChatGPT chat surface.

Strategic Positioning

The Reuters framing is right: this is an enterprise AI push. Dots is not a consumer toy — it is designed to sit inside existing workplace tools and handle ongoing work, not just answer questions. The always-on model means OpenAI is betting on recurring compute consumption per user rather than spike-and-idle query traffic.

That changes the unit economics. A dot running background tasks 24 hours a day at GPT-6 Astra token prices is a very different bill from occasional ChatGPT sessions. OpenAI has not published Dots-specific pricing beyond what’s included in existing Pro and Enterprise plans.

The competition with Muse also signals that personal agent interfaces — not raw model capability — are the next differentiation layer. Both companies are trying to claim the ambient AI position: the thing running in the background, learning your preferences, acting on your behalf.