OpenAI is preparing to combine ChatGPT’s traditional chat experience with its more advanced Work capabilities, bringing everyday conversations and multi-step AI tasks into a unified interface. The change is expected to make it easier for users to move from asking questions to having ChatGPT research information, work with files, build websites and complete projects without switching between separate modes.
However, the product merger comes with an important change to how usage is measured. Starting in December 2026, Plus, Pro and Business subscriptions are expected to share a combined usage allowance across ChatGPT and Codex. Instead of treating conversational AI and agent-based work as separate activities, OpenAI will measure consumption through tokens, with more advanced models and complex tasks using more of the available allowance. The development signals a shift away from the idea that a fixed subscription can provide unlimited access to every level of AI computing.
Key Takeaways
- OpenAI is preparing to merge Chat and Work into a more unified ChatGPT experience.
- The combined interface will allow users to move from conversations to multi-step tasks without switching modes.
- Starting in December 2026, Plus, Pro and Business plans will share usage allowances across ChatGPT and Codex.
- Usage will be measured in tokens, with more demanding models and tasks consuming more allowance.
- Instant conversations will not count towards the shared usage budget.
- Users who reach their allowance will be able to wait for a reset, upgrade or purchase additional credits for eligible features.
OpenAI Wants to Combine ChatGPT Chat and Work
ChatGPT has traditionally separated quick conversations from more advanced agentic work.
The Chat experience is designed for everyday questions, brainstorming, explanations and research. Work, meanwhile, is designed for longer tasks in which ChatGPT can use files, tools and connected services to produce completed deliverables.
OpenAI now wants to bring these experiences closer together.
According to an update from the company, a combined experience will allow users to move from exploring an idea to asking ChatGPT to carry it through. For example, a user could begin by discussing a business idea and then ask the assistant to research the market, analyse documents, create a presentation or build a website within the same workflow.
The company is initially testing the combined experience with a small group of users before expanding availability.
Codex, OpenAI’s software development environment, will remain a separate experience in the ChatGPT desktop application.
The broader objective is to make ChatGPT function as a single workspace rather than requiring users to understand and switch between multiple operating modes.
Shared Token Allowances Will Change How Subscriptions Work
The more significant development concerns subscription usage.
OpenAI’s updated Help Center documentation says that starting in December 2026, Plus, Pro and Business plans will share a combined usage allowance across ChatGPT and Codex.
This means that usage in one experience can affect the remaining allowance available in another.
For example, a subscriber who spends substantial resources on a complex coding project through Codex may have less of their shared allowance available for advanced tasks in ChatGPT Work.
The system will measure usage through tokens, which are units of information processed and generated by AI models.
The amount consumed will depend on several factors:
- The model selected by the user.
- The amount of information provided in a prompt.
- The length of the model’s response.
- The complexity of the task.
- The reasoning and computing resources required.
A short question handled by a lighter model will generally consume less of the allowance than a lengthy research project involving multiple files, tool calls and generated outputs.
What Happens When Users Reach Their Usage Limit?
OpenAI says subscribers will be able to monitor their combined remaining usage and reset times through the Settings section.
If users reach their allowance, the company identifies three ways to continue:
- Wait for the reset: Users can continue when their usage allowance becomes available again.
- Upgrade the subscription: Users can move to a plan with a higher included allowance.
- Purchase additional credits: Eligible features can be used beyond the included allowance through additional credit purchases.
This introduces a more consumption-aware approach to AI subscriptions.
Instead of treating all interactions as roughly equivalent, the system recognises that different tasks place different demands on computing infrastructure.
Instant Chat Will Remain Outside the Shared Usage Budget
There is an important exception to the new arrangement.
OpenAI says conversations using Instant models will not draw from the shared ChatGPT and Codex allowance.
This means everyday conversational use will remain separate from the token budget allocated to more advanced models and agentic tasks.
The distinction is important because not every ChatGPT interaction requires extensive reasoning or autonomous execution.
A user asking for a definition, a short explanation or a simple rewrite does not necessarily need the same computing resources as someone asking an AI agent to analyse several documents and produce a detailed business report.
OpenAI’s approach separates these lighter interactions from more resource-intensive work.
Why OpenAI Is Moving Away From Unlimited AI Usage
The shift reflects the economics of operating increasingly capable AI systems.
Traditional software subscriptions can often provide broad access because the cost of serving an additional user or request is relatively low. Advanced AI models, however, require computing resources for every interaction.
The cost can rise significantly when users request longer reasoning processes, large document analysis, code generation or autonomous agents that perform multiple actions.
As AI products become more capable, the difference between a simple prompt and a complex task becomes increasingly important for service providers.
OpenAI has also previously discussed the long-term sustainability of unlimited AI subscriptions. In March 2026, ChatGPT head Nick Turley suggested that unlimited plans may eventually become difficult to maintain as models become more powerful and expensive to operate.
The shared allowance system is consistent with that direction: users can continue receiving conversational access while more demanding computing tasks draw against a defined usage budget.
How the Change Could Affect Plus, Pro and Business Subscribers
The impact will vary depending on how subscribers use ChatGPT.
| User type | Potential impact |
|---|---|
| Casual ChatGPT users | Limited impact if most conversations use Instant models |
| Plus subscribers | Advanced Work and Codex tasks will draw from a shared allowance |
| Pro subscribers | More included usage, but complex tasks will still consume allowance |
| Business subscribers | Teams may need to monitor combined usage across workflows |
| Developers | Coding tasks through Codex will be part of the shared usage system |
For people who mainly use ChatGPT for everyday conversations, the change may be less noticeable.
However, users who regularly depend on advanced reasoning, coding agents, file analysis and long-running Work sessions may need to pay closer attention to their consumption.
OpenAI says it expects most users will not need to wait for their usage to reset more frequently than they do currently.
The Bigger Picture
The merger of Chat and Work reflects a broader transition in the AI industry. Chatbots are increasingly becoming platforms that can execute tasks rather than simply generate answers. This creates a different cost structure because an AI agent may perform multiple reasoning steps, interact with tools and process large amounts of information to complete one request.
OpenAI’s shared allowance system is an attempt to connect subscription access more closely with the computing resources consumed. It also reflects a wider industry shift towards usage-based pricing, credits and tiered access for advanced AI capabilities. The challenge will be balancing predictable subscription costs with the growing expense of agent-based computing.
Looking Ahead
The next major milestone is the planned December 2026 introduction of shared usage allowances for Plus, Pro and Business subscribers. The unified Chat-and-Work experience is also expected to expand beyond its initial testing group. Users will need to understand how their chosen models and workflows affect their remaining allowance.
For OpenAI, the long-term objective is to make ChatGPT a unified environment for conversations and completed work while keeping the cost of advanced AI services manageable. Whether subscribers accept the new arrangement will depend on how generous the included allowances are, how clearly usage is displayed and how much value users receive from the additional agentic capabilities.
Frequently Asked Questions
1. What is OpenAI changing in ChatGPT?
OpenAI is preparing to bring Chat and Work closer together, allowing users to move from ordinary conversations to multi-step tasks in a unified experience.
2. When will ChatGPT and Codex start sharing usage allowances?
OpenAI’s updated Help Center documentation says the shared allowance is expected to begin in December 2026 for Plus, Pro and Business plans.
3. Will normal ChatGPT conversations count towards the shared limit?
Instant conversations will not draw from the shared usage allowance. More advanced models and complex tasks will consume the available budget.
4. How will OpenAI measure AI usage?
Usage will be measured in tokens, which represent the information processed and generated by AI models. Consumption will depend on the model and the complexity of the task.
5. What happens when a subscriber reaches the usage limit?
Users can wait for the allowance to reset, upgrade their subscription or purchase additional credits for eligible features.
6. Will Codex remain a separate product?
Codex will remain a separate experience in the ChatGPT desktop application, even as its usage allowance becomes connected to ChatGPT.
7. Does this mean ChatGPT is ending unlimited access completely?
No. Instant conversations will remain outside the shared usage allowance. The change primarily affects advanced models and more demanding Work and Codex tasks.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
