Cursor is increasing the amount of included usage available to subscribers across its plans, giving developers more access to its first-party Cursor Models, including Auto, from August 24, 2026. The company informed users about the change through an email, saying the higher included limits will apply automatically to their current billing cycle, with no action required from subscribers.
The announcement comes as Cursor continues shifting its product toward agentic software development, where developers increasingly rely on AI models for multi-step coding, debugging, testing and repository-level tasks. At the same time, Cursor is changing how Auto is priced, making the relationship between model selection, token usage and subscription limits more important for users.
Cursor To Increase Included Usage From August 24
According to the email sent to users, Cursor will increase the included usage limits for Cursor Models, including Auto, effective August 24.
The increase applies across the company’s plans rather than being limited to a particular subscription tier. Existing subscribers do not need to change their plans or take any action to receive the updated allowance.
Cursor Usage Update At A Glance
| Parameter | Details |
|---|---|
| Company | Cursor |
| Change | Higher included usage limits |
| Effective date | August 24, 2026 |
| Coverage | All plans |
| Models covered | Cursor Models, including Auto |
| Existing billing cycle | Higher limits apply |
| User action required | No |
| Auto pricing | Also being updated |
| Main objective | More included AI coding usage |
Cursor’s current documentation separates its Cursor Models usage pool from its Other Models pool. The Cursor Models pool currently covers first-party models including Cursor Grok 4.6, Grok 4.5 and Composer 2.5.
What Is Changing With Auto?
The usage-limit increase comes alongside a change to Auto pricing.
Cursor’s user email says Auto pricing will now be based on the model each request is routed to rather than one flat rate. The company said this change is intended to allow it to continue routing requests toward cost-efficient models for everyday development.
Auto itself is not being removed.
Cursor says Auto will continue to automatically route requests to the model it considers appropriate for the task. The company also says the models that Auto routes to are not changing as part of this announcement.
Auto’s New Pricing Direction
Developer selects Auto
↓
Cursor evaluates request
↓
Auto routes request
↓
Usage reflects routed model
↓
Request consumes included usage
The important change is therefore how usage is measured rather than how developers select Auto.
Cursor Models Get More Included Usage
Cursor’s first-party model pool is becoming increasingly important to the company’s pricing strategy.
The current Cursor documentation lists three models in the Cursor Models pool:
| Cursor Model | Input Price* | Output Price* |
|---|---|---|
| Cursor Grok 4.6 | $2 / 1M tokens | $6 / 1M |
| Cursor Grok 4.5 | $2 / 1M tokens | $6 / 1M |
| Cursor Composer 2.5 | $0.50 / 1M tokens | $2.50 / 1M |
*Current published model rates; the August 24 announcement concerns included usage limits rather than these individual rates.
The increased allowance should give developers more room to use these models before exhausting their included monthly usage.
Auto Will Continue To Route Requests Automatically
Auto is designed to remove the need for developers to manually select a model for every task.
A simple request might be handled differently from a complex coding task, allowing Cursor to make the model-selection decision in the background.
Typical Auto Workflow
| Developer Action | Cursor Action |
|---|---|
| Select Auto | Enables automatic routing |
| Enter coding task | Analyses request |
| Submit prompt | Selects appropriate model |
| Agent works | Usage is recorded |
| Task completed | Developer reviews result |
Cursor’s documentation currently describes three Auto modes — Auto Cost, Auto Balance and Auto Intelligence — although Auto Balance and Auto Intelligence are currently described as available to Teams and Enterprise users.
For individual users, Cursor’s documentation currently lists Auto Cost at a flat token rate, while the company’s latest email indicates that Auto pricing is being updated from August 24.
Why Cursor Is Increasing Limits Now
The increase comes as developers are using AI agents for substantially more work than traditional autocomplete.
An AI coding agent can inspect an entire repository, modify multiple files, run tests, identify errors and continue iterating. Those workflows can consume considerably more model usage than a simple question-and-answer interaction.
Traditional AI Coding
Prompt
→ Code suggestion
→ Developer edits
→ Developer tests
Agentic Coding
Goal
↓
Repository analysis
↓
Planning
↓
Multi-file edits
↓
Testing
↓
Error detection
↓
Additional changes
↓
Completed task
Higher included limits can make the second workflow more practical for developers who use Cursor throughout the day.
Cursor Already Uses Separate Usage Pools
Cursor’s current pricing system uses separate pools for its own models and third-party models.
The Cursor Models pool covers models such as Grok and Composer, while the Other Models pool covers third-party models and is charged according to model API pricing.
Cursor’s Usage Structure
| Usage Pool | Models | Current Structure |
|---|---|---|
| Cursor Models | Grok 4.6, Grok 4.5, Composer 2.5 | Included usage |
| Other Models | OpenAI, Anthropic, Google and others | Separate included pool on eligible plans |
| On-demand | Additional usage | Pay-as-you-go |
This separation means the August 24 increase for Cursor Models should not automatically be interpreted as an equivalent increase in every third-party model allowance.
The user’s email specifically refers to increased included usage for Cursor Models, including Auto.
Pro, Pro+ And Ultra Already Offer Different Usage Levels
Cursor’s current plans provide different levels of access.
According to Cursor’s pricing documentation, Pro is priced at $20 per month, Pro Plus at $60 and Ultra at $200. The company describes Pro Plus as providing three times the Pro Agent limits and Ultra as providing 20 times the Pro Agent limits.
Cursor Individual Plans
| Plan | Monthly Price | Agent Positioning |
|---|---|---|
| Pro | $20 | Extended limits |
| Pro Plus | $60 | 3X Pro Agent limits |
| Ultra | $200 | 20X Pro Agent limits |
The August 24 change therefore increases the amount of included usage available within the existing plan structure rather than introducing a completely new subscription tier.
India Users Also Have The New Cursor Start Plan
The usage announcement comes shortly after Cursor launched a lower-cost plan specifically for developers in India.
Cursor Start costs ₹649 per month, including taxes, and provides access to Cursor models such as Grok and Composer along with cloud agents, iOS access and developer extensions.
Cursor Plans In India
| Plan | Price | Key Positioning |
|---|---|---|
| Hobby | Free | Limited AI usage |
| Start | ₹649/month | Everyday agentic development |
| Pro | ₹1,800/month* | Broader model access |
| Pro+ | Higher | Heavy AI usage |
| Ultra | Higher | Power users |
*Cursor’s India-facing pricing page currently displays Pro at ₹1,800/month, while its global documentation lists the plan at $20/month; taxes and billing arrangements can affect displayed pricing.
The Start plan is particularly relevant to India’s rapidly expanding developer community. Cursor said in its July announcement that its Indian user base had grown to more than 3 million developers and that India had more power users, measured by agent requests per developer, than any other market.
The Change Could Benefit Heavy AI Users Most
The practical benefit of higher included limits will vary depending on how much AI a developer already uses.
A casual user who rarely reaches the existing allowance may notice little difference. Developers who run agents continuously, however, could gain significantly more usable capacity.
Who Benefits Most?
| User | Potential Benefit |
|---|---|
| Occasional coder | Low |
| Student | Moderate |
| Freelancer | Moderate-high |
| Daily developer | High |
| Agent-heavy developer | Very high |
| Multi-agent user | Very high |
| AI-first development team | High |
For developers building applications almost entirely through AI agents, the additional allowance could reduce the frequency with which they have to monitor usage or switch models.
More Usage Could Encourage Larger Coding Tasks
One potential effect of higher limits is that developers may become more willing to assign larger tasks to Cursor.
Instead of breaking an application update into numerous small prompts to conserve usage, developers could allow an agent to work through a broader objective.
Before Higher Limits
Large task
↓
Concern about usage
↓
Break into smaller tasks
↓
More manual supervision
After Higher Limits
Large task
↓
Give agent broader objective
↓
Agent works across files
↓
Test + iterate
↓
Developer reviews
This is closely aligned with Cursor’s broader push toward agentic development.
Cursor Is Expanding Its Agent Ecosystem
The company is not limiting its AI development strategy to the desktop editor.
Cursor now offers cloud agents that can build, test and open pull requests while developers continue working. It also provides integrations involving plugins, MCP servers, hooks and skills.
The result is a broader development workflow:
Developer
↓
Cursor IDE
↓
AI Agent
↓
Cloud Agent
↓
Build + Test
↓
Pull Request
↓
Human Review
As this workflow becomes more common, usage capacity becomes increasingly important because agents can perform many more model operations than traditional autocomplete.
There Is A Trade-Off With The Auto Pricing Change
While the higher included allowance is positive for users, the Auto pricing change introduces a more complicated calculation.
Cursor’s email says the company is moving Auto away from a single flat rate toward pricing based on the model handling each request. The company says this allows it to continue serving cost-efficient models for everyday tasks.
That means the value of the higher allowance will depend partly on the models Auto selects and how much work individual requests require.
The Key Trade-Off
Higher Included Usage
More Flexible Model-Based Pricing
but
Potentially Different Consumption Per Request
For users who rely heavily on Auto, this is likely to be the most important part of the August 24 update to understand.
Cursor Users Are Already Discussing The Change
The announcement has generated mixed reactions among Cursor users.
Some users view the higher limits positively, particularly those who primarily use Composer or Grok. Others are questioning how the new Auto pricing will affect their effective monthly usage because a request may consume usage differently depending on the model selected.
That uncertainty is understandable because the benefit of a larger quota depends on both the size of the quota and the consumption rate of the requests.
Cursor’s documentation says different models have different API costs and that model selection affects how quickly included usage is consumed.
What Cursor Users Should Know Before August 24
For existing subscribers, the main points are straightforward.
August 24 Checklist
- Included usage limits will increase.
- The increase applies to Cursor Models, including Auto.
- The change takes effect August 24, 2026.
- Users do not need to manually activate the increase.
- Cursor is also changing how Auto is priced.
- Auto will continue to perform automatic model routing.
- Third-party model usage remains a separate consideration.
- Users can monitor consumption through Cursor’s usage dashboard.
The Bigger Picture
Cursor’s decision to increase included usage limits across plans reflects the rapid shift from traditional AI-assisted programming toward agentic development. Developers are increasingly asking AI systems to handle complete engineering tasks rather than simply generate individual lines of code. Higher limits give those workflows more room to operate within a subscription.
The timing is also significant because Cursor is simultaneously refining its model and Auto pricing. The company is effectively giving users more included capacity while changing how Auto requests are priced. That means the real impact will depend not just on the headline increase in limits, but also on how efficiently developers’ requests consume the expanded allowance.
Looking Ahead
The August 24 change could make Cursor more attractive to developers who use AI heavily, particularly those relying on Composer, Grok and Auto for everyday development. The larger included allowance may encourage developers to delegate longer coding and debugging tasks to agents, strengthening Cursor’s position as an AI-native development environment.
For Auto users, however, the pricing change will be just as important as the higher limits. Developers will want to monitor how their usage changes after the update and compare the effective consumption of Auto with manually selected models. If Cursor can deliver more useful agentic work without forcing users into frequent upgrades or unexpected additional usage, the change could strengthen its position in an increasingly competitive AI coding market
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
