Grok 4.6 in Cursor: Pricing and Best Uses

Cursor and SpaceXAI released Grok 4.6A Cursor and SpaceXAI model for complex coding and knowledge work, still in the Cursor model pool and succeeded by Grok 4.7; it is a selectable model and is separate from the Grok Bot product. Press Enter for the full definition. on August 12, 2026. Cursor positions it for work that has to survive many steps: researching unfamiliar material, navigating a codebase, using tools, checking results, correcting course, and turning a broad idea into a working application or document.

It is no longer the newest Grok. Cursor's Grok 4.6A Cursor and SpaceXAI model for complex coding and knowledge work, still in the Cursor model pool and succeeded by Grok 4.7; it is a selectable model and is separate from the Grok Bot product. Press Enter for the full definition. page now recommends Grok 4.7Cursor and SpaceXAI's current lead Cursor model, in the weekly pool at $2 per million input tokens and $6 per million output, with a 256k context window and a 500k variant; above 256k input tokens it bills at 2x the standard rate. Press Enter for the full definition., its successor, for stronger self-verification and longer-running work. Grok 4.6 remains available, at the same standard rates as Grok 4.7: $2 input and $6 output per million tokens. For the wider choice, start from the current model guide.

The documented gains over Grok 4.5 are better instruction following, stronger long-horizon work and better first passes on visual and interactive projects. Use them to decide where to try it first, and test it on work whose correct result you can check.

Don't confuse the model, which you select inside Cursor's agent surfaces, with Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition.. Grok Bot is a separate product that works on its own cloud computer, runs routines and has its own plan access. BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. is different again: it is Cursor's pull request reviewer.

Cursor says Grok 4.6A Cursor and SpaceXAI model for complex coding and knowledge work, still in the Cursor model pool and succeeded by Grok 4.7; it is a selectable model and is separate from the Grok Bot product. Press Enter for the full definition. stays with complex tasks across many steps and follows detailed instructions more reliably than Grok 4.5. Its launch post also reports stronger first passes on visual and interactive projects, and more self-testing and verification during longer runs. These gains matter most on long tasks, where a run that loses track of the goal costs more than one extra model turn.

The launch post credits the change to a longer supplemental training run, supervised trajectories regenerated across effort levels and agent harnesses, and reinforcement learning across general coding, knowledge work, kernel optimization, web development and computer-aided design. That explains how Cursor trained the behavior. Whether it beats an alternative on your stack is something only your own tests can show.

Cursor's launch post also cites benchmark results, including that Grok 4.6A Cursor and SpaceXAI model for complex coding and knowledge work, still in the Cursor model pool and succeeded by Grok 4.7; it is a selectable model and is separate from the Grok Bot product. Press Enter for the full definition. matches GPT-5.6 Sol on one composite intelligence index. Treat that as launch evidence to check against a repository test. A benchmark can't see your build system, your review standard or your company policy, and it can't tell whether your task needs visual judgment or code reasoning.

You can set Grok 4.6A Cursor and SpaceXAI model for complex coding and knowledge work, still in the Cursor model pool and succeeded by Grok 4.7; it is a selectable model and is separate from the Grok Bot product. Press Enter for the full definition. to low, medium, high or xhigh effort. High is the default on plans that let you choose. A higher setting gives the model more time to work through a hard task. The per-token rate stays the same, but a deeper run can use more tokens and take longer.

Inside Cursor, the model can search files and the web, read and edit files, run shell commands, control a browser, take screenshots, generate images, fetch rules and ask clarifying questions. Cursor also documents image reading for vision-capable models. What a given run can use still depends on the product surface, approval mode, sandbox, team policy and the credentials it has.

The longer the tool list, the more a narrow task contract is worth. Name the files or systems in scope, the commands that prove the work is done, and the actions that need approval. The model can check its own work, but only your acceptance test tells it what finished means in this repository.

Included usage comes first: Grok 4.6A Cursor and SpaceXAI model for complex coding and knowledge work, still in the Cursor model pool and succeeded by Grok 4.7; it is a selectable model and is separate from the Grok Bot product. Press Enter for the full definition. draws on the pool for Cursor models that your plan includes. Beyond the pool, Cursor publishes Standard and Fast rates per million tokens. Its table shows no cache-write rate for either tier, and Cursor doesn't say whether the blank means cache writes are free.

Fast doubles every published rate, and it is the default speed tier on Premium and higher and on the legacy Pro, Pro+ and Ultra plans. Confirm the speed tier before cost-sensitive work starts. Standard suits queued or background work where nobody waits on each response, while Fast is easier to justify in an active pairing loop.

Cursor's launch promotion is over. For one week from August 12, the model page listed a 50% discount and the launch post described two times the included usage.

Yes, at fixed settings. On Plus and on Cursor StartAn India-only legacy individual plan at ₹649 a month centered on Cursor's own models, with cloud and mobile access but without Other Models, Auto, Bugbot, Automations or the Agent SDK; Cursor names Plus as the comparable plan on The X Subscription. Press Enter for the full definition., Grok 4.6A Cursor and SpaceXAI model for complex coding and knowledge work, still in the Cursor model pool and succeeded by Grok 4.7; it is a selectable model and is separate from the Grok Bot product. Press Enter for the full definition. runs at standard speed with medium effort, alongside Grok 4.7Cursor and SpaceXAI's current lead Cursor model, in the weekly pool at $2 per million input tokens and $6 per million output, with a 256k context window and a 500k variant; above 256k input tokens it bills at 2x the standard rate. Press Enter for the full definition., Grok 4.5 and Composer 2.5The current Composer release, better at long-running tasks and at judging when a job needs a light touch versus deep work. Press Enter for the full definition.. Start is the plan Cursor sold to developers in India before The X SubscriptionThe subscription Cursor is sold through: one plan for Grok, Cursor and X, with Free, Plus, Premium, Super and Ultra tiers, per-seat Teams, a weekly included pool for Cursor models shared with Grok, and third-party models billed at API list price from prepaid credit. Press Enter for the full definition., now legacy, at ₹649 a month with tax included.

Start also excludes BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition., Automations and the Cursor SDK. The Grok 4.6A Cursor and SpaceXAI model for complex coding and knowledge work, still in the Cursor model pool and succeeded by Grok 4.7; it is a selectable model and is separate from the Grok Bot product. Press Enter for the full definition. it includes is the same model with its speed and effort controls fixed, and AI features pause when the monthly pool runs out. For Fast, effort choice, Bugbot or Automations, move to Premium or above on The X SubscriptionThe subscription Cursor is sold through: one plan for Grok, Cursor and X, with Free, Plus, Premium, Super and Ultra tiers, per-seat Teams, a weekly included pool for Cursor models shared with Grok, and third-party models billed at API list price from prepaid credit. Press Enter for the full definition., or to legacy Pro, Pro+ or Ultra. Cursor's pricing pages don't say which plans on The X Subscription include the SDK.

Cursor publishes no size for Start's allowance. Its help page says the monthly pool is larger than the free Hobby plan's, and that the Grok models use it faster than Composer 2.5The current Composer release, better at long-running tasks and at judging when a job needs a light touch versus deep work. Press Enter for the full definition. does. Check the dashboard for the balance you have left.

Start with model access, the usage pool and Privacy ModeCursor's setting that guarantees code data is not used for training by Cursor or its model providers, and that an admin can enforce org-wide; data-retention terms are a separate, contractual layer. Press Enter for the full definition.. Cursor says that with Privacy Mode on, code is never used for training by Cursor or its model providers. Its privacy page adds that most models run under zero data retention agreements, and names the Claude Fable models as exceptions. It publishes no retention period specific to Grok 4.6A Cursor and SpaceXAI model for complex coding and knowledge work, still in the Cursor model pool and succeeded by Grok 4.7; it is a selectable model and is separate from the Grok Bot product. Press Enter for the full definition..

Data residency is a separate check. Cursor's Models & Pricing page marks Grok 4.6A Cursor and SpaceXAI model for complex coding and knowledge work, still in the Cursor model pool and succeeded by Grok 4.7; it is a selectable model and is separate from the Grok Bot product. Press Enter for the full definition. as eligible for US-only data residency. That option is for Enterprise customers, is enabled per team, and adds 10% to the price of eligible models.

Keep Grok 4.6A Cursor and SpaceXAI model for complex coding and knowledge work, still in the Cursor model pool and succeeded by Grok 4.7; it is a selectable model and is separate from the Grok Bot product. Press Enter for the full definition. allowed after a move to Grok 4.7Cursor and SpaceXAI's current lead Cursor model, in the weekly pool at $2 per million input tokens and $6 per million output, with a 256k context window and a 500k variant; above 256k input tokens it bills at 2x the standard rate. Press Enter for the full definition., because Cursor Router requires it and blocking it disables the router.

At launch, Cursor also made Grok 4.6A Cursor and SpaceXAI model for complex coding and knowledge work, still in the Cursor model pool and succeeded by Grok 4.7; it is a selectable model and is separate from the Grok Bot product. Press Enter for the full definition. available in Grok Build, the SpaceXAI API, OpenRouter, Vercel and Cloudflare. Availability there doesn't carry over Cursor's billing, tool access, retention or approval controls.

1. 1Confirm the model is allowed for the relevant team or group, and keep it allowed if you use Cursor Router.
2. 2Choose Standard or Fast, and set an effort level that matches the task.
3. 3Run a representative task with fixed acceptance criteria and record the evidence.
4. 4Check Privacy ModeCursor's setting that guarantees code data is not used for training by Cursor or its model providers, and that an admin can enforce org-wide; data-retention terms are a separate, contractual layer. Press Enter for the full definition., retention exceptions and residency eligibility against current documentation.
5. 5Review usage after one billing cycle before you make it the default everywhere.

Keep a dated record of what you verified, when, and on which task. Grok 4.6A Cursor and SpaceXAI model for complex coding and knowledge work, still in the Cursor model pool and succeeded by Grok 4.7; it is a selectable model and is separate from the Grok Bot product. Press Enter for the full definition. shows why: since launch, it has stopped being the Grok model Cursor recommends, become eligible for US-only data residency and become the model Cursor Router requires.