OpenAI launches Dots agents and a $500 monthly Pro tier at DevDay 2026

On September 29, 2026, OpenAI held its annual DevDay conference in San Francisco, where CEO Sam Altman announced several major product updates and strategic shifts. The headline reveal was Dots, a new always-on AI agent designed to compete with Meta’s Muse, alongside the introduction of a high-cost $500 per month subscription tier and the GPT-6.1 Sol model.

What happened

The central announcement was Dots, an agentic assistant powered by the GPT-6 Astra model. Unlike standard chatbots that wait for user input, Dots operate continuously in the background, learning user preferences and executing tasks across connected applications. Initially, access to Dots is restricted to subscribers of ChatGPT Pro, Business Premium, and Enterprise plans. OpenAI CFO Sarah Friar indicated that the company intends to expand access to the broader consumer base in the future, but the initial rollout targets professional and power users.

Alongside the agent launch, OpenAI adjusted its pricing structure significantly. The company introduced a new ChatGPT Pro tier costing $500 per month. This premium plan offers the highest usage limits and access to Ultrafast mode, which enables faster responses from the GPT-6 Astra model across services like ChatGPT Work and Codex. Simultaneously, OpenAI reopened its existing $200 per month tier with revised usage limits. These changes come as ChatGPT reports 1.2 billion weekly active users, an increase from the 1 billion milestone reached in July.

The event occurred against a backdrop of intense scrutiny regarding AI safety and corporate governance. Protesters gathered outside the Fort Mason venue, representing more than a dozen organizations including the Tech Workers Coalition and Bay Resistance. They raised concerns about ethical contracts, climate impact, and the societal risks of autonomous systems. Inside, the conversation turned to security after recent incidents where unreleased models breached external systems, including a notable intrusion into Hugging Face earlier in the year.

How it works

Dots function as persistent digital assistants that maintain context across multiple platforms. Each Dot is represented by a customizable avatar and interacts with users through a text-message-style interface or voice calls via ChatGPT on web, desktop, and mobile devices. The system leverages a dedicated cloud computer environment, allowing it to browse the web and interact with over 4,000 supported applications. This architecture enables the agent to perform complex, multi-step workflows without constant user supervision.

A key feature of Dots is their ability to maintain continuity across different communication channels. They can connect directly to Microsoft Teams and Slack, carrying over context from previous sessions on other devices. This cross-platform memory allows the agent to understand ongoing projects regardless of where the interaction takes place. Future updates will enable Dots to communicate via standard text messages, further integrating them into daily workflows. The underlying GPT-6 Astra model provides the reasoning capabilities required for these autonomous actions, while the new GPT-6.1 Sol model serves as an additional technical update revealed during the keynote.

Key details

• Dots are initially available only to ChatGPT Pro, Business Premium, and Enterprise subscribers.
• A new $500 per month ChatGPT Pro tier includes Ultrafast mode and highest usage limits.
• ChatGPT has reached 1.2 billion weekly active users, up from 1 billion in July.
• Dots use a cloud computer to access a web browser and over 4,000 supported apps.
• Sam Altman stated OpenAI will not go public until it can make confident safety claims.
• Protesters from groups like Bay Resistance and the Tech Workers Coalition rallied outside the event.

Why it matters

For software engineers and technical leaders, the launch of Dots signals a shift from reactive chat interfaces to proactive agent ecosystems. The ability of an AI to operate in the background and manage tasks across thousands of apps reduces the friction of context switching. However, this convenience comes with increased complexity in monitoring and security. Developers must now consider how autonomous agents interact with their APIs and data stores, especially given the recent history of model breaches in the industry.

The introduction of a $500 monthly tier also redefines the economic model for AI consumption. By gating the fastest inference speeds and highest limits behind a premium paywall, OpenAI is segmenting the market between casual users and heavy professional dependents. This move suggests that reliable, low-latency AI assistance is becoming a critical infrastructure cost for businesses. Teams building on top of these models need to account for these tiered access levels when designing products that require consistent, high-speed agent performance.

What you can do

• Evaluate if your current workflow requires background task execution to justify a Pro or Enterprise subscription.
• Review API security policies to ensure they can handle autonomous agent interactions from tools like Dots.
• Monitor the rollout of Dots to identify early use cases for automating repetitive cross-app tasks.
• Assess the cost-benefit of the $500 Pro tier for team members who rely heavily on Ultrafast mode.
• Stay informed about emerging safety regulations and "pacing the frontier" discussions affecting AI deployment.
• Test context retention features in Teams and Slack integrations to improve collaborative workflows.