OpenAI DevDay 2026: Dots, GPT-6.1 Sol, and the Agents API Go Mainstream

On September 29, 2026, OpenAI ran what it calls its biggest DevDay yet: more than 20 announcements spanning ChatGPT, Codex, its models, and a new way of working with agents, dropped in front of 1.2 billion weekly ChatGPT users. The interesting part for engineering teams is not any single feature but three shifts landing at once. Always-on agents called Dots move execution from "you prompt once" to "it keeps working." GPT-6.1 Sol pushes near-Astra intelligence down to a fifth of the price. And the Agents API and Decisions API turn "an agent that clicks a mouse" and "a bounded decision" into platform primitives. Here is what actually shipped, section by section, plus a checklist you can run on Monday.

1. What DevDay 2026 actually shipped

Start with the facts. According to OpenAI's official DevDay 2026 Recap, this was its biggest DevDay yet: more than 20 announcements spanning ChatGPT, Codex, its models and entirely new forms of working with AI, in front of 1.2 billion weekly ChatGPT users. The recap groups them into new ways of working with AI (Dots), developer tools in Codex and the API, ChatGPT plugin customization, improved human-AI collaboration, and subscription benefits. For engineering teams, three signals matter: agents stop being something you call and start being something that stays running; frontier-grade intelligence gets much cheaper; and both computer use and bounded decision-making become first-class API surfaces.

# DevDay 2026, in one object. OpenAI framed this as its biggest DevDay
# yet, with 20+ announcements spanning ChatGPT, Codex, models and a new
# way of working with agents. The audience number is the reason the
# distribution story matters as much as the model story.

DEVDAY_2026 = {
    "date": "2026-09-29",
    "announcement_count": "20+",
    "weekly_chatgpt_users": "1.2B",
    "surfaces": ["ChatGPT", "Codex", "API", "Agents API"],
    "headline_items": [
        "Dots (always-on agents)",
        "GPT-6.1 Sol (near-Astra at 1/5 price)",
        "Ultrafast speed tier",
        "Codex in the cloud + Codex Security Cloud",
        "Decisions API",
        "Agents API with computer use",
    ],
}

20+ launches, 1.2B weekly users: distribution and execution move together

2. Dots: from prompt-once to always-working

Dots are the biggest conceptual shift. Per OpenAI's "Introducing dots," dots are always-on agents powered by GPT-6 Astra, each with its own cloud computer and browser, that learn from feedback over time, work toward your goals 24/7, and reach more than 4,000 apps through plugins. OpenAI's own example is telling: a dot watches customer feedback for recurring requests, scopes smaller fixes, builds and tests them, and hands you complete pull requests with videos attached. The control surface is the part to read closely — OpenAI says dots add safety and privacy safeguards on top of ChatGPT's protections, with built-in safeguards, access and permissions, and action review and approvals. They begin rolling out on Pro, Business Premium and Enterprise plans in eligible markets, and OpenAI previewed specialist dots with their own identity for access management, IT-provisioned hardware and deep integrations with systems of record.

# Dots are the biggest conceptual shift: an agent that keeps working
# when you are not prompting it. OpenAI says a dot is powered by GPT-6
# Astra, has its own cloud computer, learns from feedback over time,
# and can reach 4,000+ apps through plugins.

def dot_capabilities():
    return {
        "brain": "GPT-6 Astra",
        "runtime": "its own cloud computer + browser",
        "reach": "4,000+ apps via plugins",
        "channels": ["ChatGPT (desktop/web/mobile)", "Slack", "Teams"],
        "availability": ["Pro", "Business Premium", "Enterprise"],
    }

# The control surface is the part to read carefully. Dots ship with
# built-in safeguards, access and permissions, and action review and
# approvals -- because an always-on agent with a cloud computer is also
# an always-on identity that needs boundaries.

3. GPT-6.1 Sol: near-Astra intelligence at a fifth of the price

The second shift is price. Per OpenAI's "Introducing GPT-6.1 Sol," it is an upgrade to GPT-6 Sol that nearly matches GPT-6 Astra on agentic coding, computer use and professional work at one-fifth of Astra's standard token price. The published API prices are $2 per million input tokens, $0.10 per million cached input tokens and $10 per million output tokens; cached input is roughly 95% cheaper than standard input and 50% cheaper than GPT-6 Sol's cached input. On DeepSWE v1.1, OpenAI says GPT-6.1 Sol matches GPT-6 Astra at roughly one-fifth the cost while beating GPT-6 Sol's best score by 6.4 percentage points at lower reasoning effort and cost. On OSWorld 2.0's offline computer-use set, it outperforms GPT-6 Sol by seven points at maximum reasoning effort at less than half the cost. For teams running long-horizon agents, the cached-input price is the real lever.

# GPT-6.1 Sol is the cost story. OpenAI positions it as an upgrade to
# GPT-6 Sol that nearly matches GPT-6 Astra on agentic coding, computer
# use and professional work at one-fifth of Astra's standard token price.

PRICING_PER_MILLION_TOKENS = {
    "gpt-6-astra":  {"input": 10.00, "output": 50.00, "cached_input": 1.00},
    "gpt-6.1-sol": {"input":  2.00, "output": 10.00, "cached_input": 0.10},
    "gpt-6-luna":  {"input":  0.10, "output":  0.50, "cached_input": 0.01},
}

def cost(model, input_tokens, output_tokens, cached_tokens=0):
    p = PRICING_PER_MILLION_TOKENS[model]
    return (input_tokens * p["input"]
            + output_tokens * p["output"]
            + cached_tokens * p["cached_input"]) / 1_000_000

# The cached-input price -- $0.10 per million tokens -- is the lever for
# long-running agents that reuse the same context across turns.

GPT-6.1 Sol brings near-Astra intelligence into a smaller budget

4. The developer surface: an Agents API that clicks, a Decisions API that routes

The third shift is in the API. Per the recap, the Agents API now supports computer use, so developers can build agents that interact with software to complete tasks; it also brings Codex's multi-agent capabilities, tool search, tool calling and context compaction into your application, with OpenAI running the underlying infrastructure. The new Decisions API focuses Luna on a set of user-defined questions with finite predefined answers, takes text or images, and returns answers you can use to classify content, route requests or choose an agent's next action — in limited preview, with a broad release planned. Codex moved too: cloud environments with reusable dev containers, a refreshed CLI with voice and an /agents view, a desktop code-review flow, and Codex Security Cloud for scheduled repository scans.

# The developer surface is where the DevDay news actually lands. Two
# API additions matter most for teams building agents.

AGENT_API_SURFACES = {
    "Agents API": {
        "new": "computer use (agents that click, type and read screens)",
        "also": ["multi-agent", "tool search", "tool calling", "context compaction"],
        "infra": "OpenAI runs the underlying infrastructure",
    },
    "Decisions API": {
        "job": "bounded, real-time decisions",
        "input": ["text", "images"],
        "output": "typed answers over a fixed set of options",
        "uses": ["classification", "routing", "choosing the next agent action"],
        "model": "Luna",
    },
}

# Codex also moved: cloud environments with reusable dev containers, a
# refreshed CLI with voice and an /agents view, a desktop code-review
# flow, and Codex Security Cloud that scans repos on a schedule.

5. Distribution and control: Bedrock, plugins, MCP Events and shared identity

Capabilities only matter if they reach production, and the plumbing is where DevDay spent real effort. OpenAI worked with Amazon on Bedrock Managed Agents, which takes the core capabilities of the Agents API and customises them to run natively in AWS and integrate with AWS resources, so you can build OpenAI agents that run entirely in AWS. On the plugin side, plugin extensions let developers give a plugin a home in the ChatGPT sidebar and build interactive panels; OpenAI also added support for the proposed MCP Events specification, so plugins can start automations when something happens in a connected app. For identity and billing, Sign in with ChatGPT lets you use your allowance across 16 partners — including Cognition's Devin, Notion, Vercel, T3, OpenClaw and Dactyl — alongside a new Pro 500 tier and the OpenAI Marketplace.

# Bedrock Managed Agents is the tell about how OpenAI wants agents
# deployed inside enterprises that already run on AWS.

BEDROCK_MANAGED_AGENTS = {
    "what": "Agents API core capabilities, customised to run natively in AWS",
    "integrates_with": "AWS resources",
    "where_it_runs": "entirely in AWS",
}

def route_by_residency(data_class):
    """Pick a runtime based on where the data is allowed to live."""
    return {
        "public": "OpenAI-hosted Agents API",
        "aws_resident": "Bedrock Managed Agents, powered by OpenAI",
    }.get(data_class, "escalate to a human")

# Control plane additions to wire in the same sprint: plugin extensions,
# the proposed MCP Events spec for event-triggered automations, and
# Sign in with ChatGPT (16 launch partners) for shared billing identity.

Bedrock, plugins and MCP Events decide how far capability travels

6. A Monday checklist

Turn the facts into work. First, pilot a single dot to learn what an always-on agent needs: an explicit scope and approval gates for sensitive actions. Second, model the savings of moving workloads from Astra to GPT-6.1 Sol — at published prices, most "good enough" agentic coding and computer-use tasks should cost about 80% less in tokens, freeing Astra for the hardest scientific and research work. Third, if you already build on the Agents API, trial computer use and context compaction first, and evaluate the Decisions API for high-frequency, low-latency routing. Fourth, if data must stay inside AWS, look at Bedrock Managed Agents. Fifth, stand up identity and approvals before you widen autonomy: dot action review, per-user plugin access, and MCP Events-triggered automations all deserve a least-privilege and audit pass — which is exactly where tools like our API Rate Limit Calculator and API Key Rotator earn their keep.