OpenAI Releases GPT-5.6 and ChatGPT Work: A Complete Enterprise AI Stack

30-Second Summary

OpenAI released three new models under the GPT-5.6 family on July 9, 2026: Luna for speed, Terra for balanced performance, and Sol for flagship-grade intelligence. The same day, OpenAI launched ChatGPT Work, an autonomous agent product that takes a defined business outcome, connects to your tools, and independently completes complex, multi-step jobs over hours, returning finished spreadsheets, slides, documents, and web portals. Together, the releases mark OpenAI's clearest move to date from AI assistant to AI operator.

---

At a Glance

• Topic: Enterprise AI
• Company: OpenAI
• Date: July 9, 2026
• Announcement: GPT-5.6 model family (Luna, Terra, Sol) released to general availability across ChatGPT, the API, Codex, and GitHub Copilot, alongside ChatGPT Work, a new autonomous agent product for Pro, Enterprise, and Edu users
• What Changed: OpenAI moved from a single frontier model to a three-tier architecture with clear cost and capability tradeoffs, and launched a separate product that executes multi-step business workflows autonomously rather than responding to conversational prompts
• Why It Matters: Enterprise operators now have access to a tiered model portfolio suited to different task types and budgets, plus an agent product with scheduled automation, connected tool access, and enterprise-grade governance controls
• Who Should Care: Business operators and team leads using ChatGPT Business or Enterprise, developers building on the OpenAI API, and anyone whose team produces recurring operational deliverables

---

Key Facts

• GPT-5.6 Sol: $5 per million input tokens, $30 per million output tokens
• GPT-5.6 Terra: $2.50 per million input tokens, $15 per million output tokens
• GPT-5.6 Luna: $1 per million input tokens, $6 per million output tokens
• All three variants share a 1,050,000-token context window and 128,000 maximum output tokens
• Knowledge cutoff: February 2026
• Sol includes an ultra mode that delegates sub-tasks to specialist submodels for complex reasoning
• GPT-5.6 introduces programmatic tool calling, explicit prompt caching breakpoints, persisted reasoning, and multi-agent orchestration in beta via the Responses API
• ChatGPT Work is available on web and mobile for Pro, Enterprise, and Edu users immediately, with Plus and Business access rolling out shortly
• ChatGPT Work includes Scheduled Tasks for recurring automation and Sites for converting work outputs into shareable interactive web apps
• Enterprise admins get centralised controls over tool access and an auto-review gate that evaluates consequential actions before execution

---

What Happened

OpenAI released GPT-5.6 on July 9, 2026, ending a limited government-restricted preview period that had begun on June 26. The model arrives as a family of three variants rather than a single flagship: Luna for speed and cost efficiency, Terra as a balanced mid-tier option, and Sol as the flagship for complex reasoning and knowledge work.

The three-tier structure is a deliberate architecture decision. OpenAI has positioned each tier against a specific set of use cases: Luna for high-volume tasks where cost and latency matter, Terra for everyday professional work where quality and cost must balance, and Sol for the most demanding tasks where getting it right matters more than getting it fast. Sol also includes an ultra mode that routes sub-tasks to specialist submodels, a feature designed for lengthy, structured work that benefits from decomposition.

On the same day, OpenAI launched ChatGPT Work, a distinct product that sits alongside the familiar conversational interface. ChatGPT Work accepts a defined outcome, connects to a user's tools and integrations, plans a sequence of steps, and executes them independently over an extended session, sometimes hours, before returning finished material. The output format is not a chat response: it is a completed spreadsheet, slide deck, document, or web application.

ChatGPT Work also introduced two supporting features. Scheduled Tasks enables recurring automation by connecting to Slack and Microsoft Teams and turning new messages into updated documents or distribution summaries automatically. Sites converts finished work into shareable interactive portals, usable as internal dashboards, client-facing reports, project trackers, or product prototypes without any separate development work required.

---

Why It Matters

The model tier system changes the cost equation for volume work. Luna at $1 input and $6 output per million tokens is the cheapest viable frontier-class model on the market for routine business tasks. For any operation running classification, extraction, tagging, or summarisation at volume, the tier system makes GPT-5.6 worth a direct cost comparison against the current stack.

ChatGPT Work closes the gap between AI and operational staff for specific task types. The product is not better at answering questions. It is designed to remove the human from a specific category of work: multi-step tasks that require gathering information from multiple sources, transforming it, and distributing the result. For that category, ChatGPT Work replaces execution time, not just assistance.

Sites removes the developer dependency from operational reporting. Client dashboards, internal portals, and management reporting tools have historically required a developer to build and maintain. The Sites feature converts finished work outputs directly into interactive web applications, which means a business analyst can now publish a live dashboard without writing code or raising a development ticket.

Multi-agent orchestration in beta is the signal to watch. The Responses API now supports multi-agent orchestration in beta, meaning developers can build systems where multiple GPT-5.6 instances coordinate on a task. For businesses building internal AI infrastructure, this is the feature that determines whether OpenAI's stack is viable for production agentic systems.

Enterprise governance is now built into the agent layer. ChatGPT Work includes centralised admin controls and an auto-review mechanism that evaluates consequential actions before they execute. For operators concerned about autonomous agents taking irreversible actions in connected systems, this governance layer is the difference between a controlled deployment and an uncontrolled one.

The combined release accelerates the timeline for AI-first operations. Six months ago, a business operator considering AI workflow automation faced a fragmented stack: a language model from one vendor, an agent framework from another, an integration layer from a third. GPT-5.6 with ChatGPT Work, enterprise controls, Scheduled Tasks, and Sites is a vertically integrated alternative to that fragmented stack.

---

The David and Goliath View

The release matters less for the model performance numbers and more for what it signals about OpenAI's strategic intent. ChatGPT Work is not a feature. It is a product category, and OpenAI has been watching Microsoft Copilot, Anthropic's Claude Cowork, and Google's Workspace AI compete for the enterprise workflow automation space for six months. The July 9 release is OpenAI's response, and it is more complete than most analysts expected.

For a small business operator, the honest assessment is this: ChatGPT Work will be genuinely useful for a specific slice of your operations, and genuinely irrelevant for the rest. The tasks where it delivers are defined and narrow: recurring deliverables with a predictable structure, clear data inputs, and a finished-document output. Anything requiring genuine judgement, client relationship management, or real-time human context is outside its operating range today.

The Sites feature deserves more attention than it has received. The ability to convert a finished work output into a shareable interactive web application without a developer is not a productivity upgrade for knowledge workers. It is an infrastructure shift for how small businesses build and maintain client-facing reporting and internal operational dashboards. If that sounds abstract, consider: every business that currently spends time emailing PDFs or maintaining a shared spreadsheet has a candidate for a Sites replacement.

---

Where This Fits in the AI Stack

GPT-5.6 and ChatGPT Work sit in the execution layer of the AI stack. The models handle language understanding and generation. ChatGPT Work adds an orchestration layer that plans, sequences, and executes multi-step tasks. Sites handles distribution of finished outputs to stakeholders. Together they cover the full workflow from instruction to finished deliverable, without requiring a separate agent framework, integration layer, or web development capability.

For businesses already running on OpenAI's API, GPT-5.6 is a direct upgrade path with a cleaner cost structure. For businesses using ChatGPT Business or Enterprise through the interface, ChatGPT Work is available in the same product they already use, which removes adoption friction.

---

Questions Operators Are Asking

Which GPT-5.6 tier should I use for my business? Use Luna for any task you run at volume where a small reduction in quality is acceptable: classification, summarisation, tagging, first-pass drafts. Use Terra for most professional knowledge work where quality matters but cost is a consideration. Use Sol for your most consequential tasks, strategic analysis, complex client deliverables, or anything where an error has a real cost.

Is ChatGPT Work the same as hiring an assistant? No, and the distinction matters. ChatGPT Work executes defined, structured workflows autonomously. A human assistant exercises judgement, manages relationships, and handles ambiguity. ChatGPT Work is better than a human assistant at executing a well-defined recurring process. It is not a substitute for judgement, client interaction, or work that requires reading context outside the task brief.

How do the enterprise governance controls work? Enterprise admins configure which tools ChatGPT Work can access, what context it can use, and who can invoke it. The auto-review mechanism evaluates actions involving connected tools and APIs before they execute. This means a mis-specified task is less likely to take an irreversible action in a connected system. The controls mirror what you would expect from enterprise software: centralised access management and an audit trail.

Can I use ChatGPT Work if I am on a Plus or Business plan? Not at launch. The July 9 release covers Pro, Enterprise, and Edu users. OpenAI has indicated Plus and Business access is coming shortly. If you are on a Plus plan, this is a reason to evaluate whether an Enterprise or Pro plan makes sense for your team now.

What is the multi-agent orchestration beta and should I care about it? Multi-agent orchestration in the Responses API allows developers to build systems where multiple GPT-5.6 instances coordinate on a task, each handling a component before passing results to the next agent. For most business operators using ChatGPT directly, this is background infrastructure that OpenAI is building. For developers building custom internal AI tools, it is a significant capability that makes production agentic systems viable on OpenAI's stack without a separate agent framework.

---

Citable Summary

OpenAI released GPT-5.6 on July 9, 2026, as a three-tier model family: Luna at $1/$6, Terra at $2.50/$15, and Sol at $5/$30 per million input/output tokens, all with a 1.05 million token context window. The same day, OpenAI launched ChatGPT Work, an autonomous agent product that takes a business outcome and independently executes multi-step workflows across connected tools over hours, returning finished deliverables rather than chat responses. ChatGPT Work is available to Pro, Enterprise, and Edu users immediately, with Scheduled Tasks for recurring automation and Sites for converting finished work into shareable interactive web applications. Enterprise governance controls include centralised admin management and an auto-review gate on consequential actions.