OpenAI DevDay 2026: Dots, Sol, Ultrafast

OpenAI's DevDay 2026 opening keynote in San Francisco is over, and the company published a same-day recap of more than 20 launches across ChatGPT, Codex, models, and APIs. The headline consumer product is Dots — always-on agents powered by GPT-6 Astra. The headline developer model is GPT-6.1 Sol, pitched as near-Astra intelligence at one-fifth of Astra's standard API token prices. A new Pro 500 plan and Ultrafast speed tier sit on top for users who want more usage and lower latency.

Dots: always-on agents with their own cloud computer

OpenAI's official Introducing dots post describes Dots as "remarkably capable, always-on agents" that work toward a user's goals around the clock. They are powered by GPT-6 Astra, get their own cloud computer and browser, learn from feedback over time, and can connect to more than 4,000 apps through OpenAI's plugin ecosystem. Users can message or call a Dot in ChatGPT, Slack, or Microsoft Teams, with texting listed as coming soon.

Availability, per OpenAI: Dots are rolling out in ChatGPT to Pro and Business Premium users in eligible markets. The first Dot is included in those plans at no extra cost, with an allowance for deeper work and extended limits for the first month after launch. Enterprise, Edu, and Healthcare workspaces can try a beta when an admin enables it (off by default). OpenAI also previewed specialist dots for organizations — separate identity, credentials, and system access — plus work with Microsoft to bring specialist dots into Agent 365 governance.

GPT-6.1 Sol: near-Astra coding and computer use, lower API price

One week after GPT-6 Sol and Luna, OpenAI introduced GPT-6.1 Sol as a major upgrade focused on agentic coding, computer use, and professional work. The company says it approaches GPT-6 Astra's intelligence on those workloads at one-fifth of Astra's standard input and output token prices. On DeepSWE v1.1, OpenAI reports GPT-6.1 Sol matching Astra at roughly one-fifth the cost while beating GPT-6 Sol by 6.4 percentage points at a lower reasoning effort.

API pricing from OpenAI's GPT-6.1 Sol page: $2 per million input tokens, $0.10 per million cached input tokens, and $10 per million output tokens, under the model id gpt-6.1-sol. Cached input is described as 95% less than standard input pricing and 50% less than GPT-6 Sol's cached input pricing. Availability: ChatGPT Work and Codex for Plus, Pro, Business, Enterprise, and Edu starting today; API access for developers; not yet available in Chat. GPT-6.1 Sol Ultrafast is listed as coming soon.

Ultrafast, Private Intelligence, and Pro 500

Ultrafast is OpenAI's new premium speed tier. The DevDay recap says it offers up to 8× faster token generation (300 tokens per second) in Codex and up to 6× in the API. GPT-6 Astra Ultrafast is available today in the API and in ChatGPT Work and Codex on Pro 500 and Enterprise plans; GPT-6.1 Sol Ultrafast is coming soon. CNBC's live DevDay coverage separately quoted Altman saying users "are really going to love this and not want to go back."

Private Intelligence is a privacy stack for businesses: Zero Data Retention with Private Safety Processing for automated safety reviews without OpenAI personnel accessing underlying content, plus a Private Inference preview coming this fall that combines confidential computing with verifiable controls. OpenAI says it designed the systems with large customers including Cisco, Databricks, and Snowflake.

Pro 500 is the new top consumer/pro tier. OpenAI's recap says it offers the highest usage allowance — 25 times the ChatGPT Plus allowance — and includes Ultrafast, available now. OpenAI's public recap does not spell out a dollar price on that page; CNBC confirmed the plan name and that it bundles highest usage with Ultrafast across ChatGPT and Codex.

Developer stack: Codex, Decisions API, Agents API, Bedrock

Codex expands beyond the laptop: cloud environments for running from any device (Plus through Enterprise), a refreshed CLI with voice steering and an /agents view, desktop Code Review for GitHub PRs and GitLab MRs with optional automatic cloud reviews, and Codex Security Cloud for scheduled repo scans that investigate findings and prepare fixes — including Daybreak Blue model access without a separate Daybreak app (Pro, Business, Enterprise, Edu).

On the API side, Decisions API puts Luna on finite, developer-defined multiple-choice questions for classification and routing (limited preview now, broader release in coming days). The Agents API gains computer use so agents can operate software, plus Codex-style multi-agent tools, tool search, and context compaction, with OpenAI hosting the infrastructure — available via API and in Codex and ChatGPT Work on Pro 500 and Enterprise. With Amazon, OpenAI also announced Bedrock Managed Agents so the same agent capabilities can run natively inside AWS.

ChatGPT as a workplace: plugins, Space, Pages, Slack and Teams

Plugin extensions open the same surface OpenAI uses for ChatGPT features so developers can ship sidebar homes, interactive panels, and custom file viewers. Plugin Creator, a redesigned submission flow, better discovery, Sites that host plugins, and MCP events for plugin automations round out the builder path. Users still choose which plugins to install and what each one can access.

ChatGPT Space is a shared home where teammates, ChatGPT, and a Dot collaborate on knowledge (Pro, Business, Enterprise on desktop and web; mobile can find/read/share pages now, with create/edit coming). Pages are collaborative documents for writing, research, charts, and images. Collaborative slides are coming in the coming weeks with multi-user and agent editing plus PowerPoint/Google Slides export. Business and Enterprise get teams, shared tasks, @ChatGPT in Slack and Microsoft Teams (teammates can use it without an individual ChatGPT license), and a Meetings plugin beta on macOS desktop for Pro and Business that saves notes and action items into Space while deleting the audio afterward.

Marketplace, Sign in with ChatGPT, and what this means

OpenAI Marketplace lets eligible enterprise customers apply part of an existing OpenAI commitment toward approved partner software. The first 32 partners named in the recap include Adobe and Figma; Sierra, Decagon, HubSpot, Salesforce, and ServiceNow; Harvey and Legora; Palo Alto Networks and CrowdStrike; and Baseten for open-source models. Sign in with ChatGPT extends plan usage across 16 partners (OpenAI lists Cognition's Devin, Notion, Vercel, T3, OpenClaw, and Dactyl among them), with identity available globally and plan usage for Plus and Pro in participating tools.

Read against Monday's GPT-6.1 Astra shelving, DevDay is OpenAI shipping the agent product line (Dots on Astra) and a cheaper near-Astra model (6.1 Sol) while keeping the higher-risk 6.1 Astra update off the shelf. For developers, the practical checklist is gpt-6.1-sol pricing, Ultrafast on Astra today, Agents API computer use, and Decisions API preview. For teams, it is Space, Pages, Slack/Teams @ChatGPT, and whether your admin turns on Dots.