C
Blog · Pricing · Switch and combine accounts Today we’re introducing The X Subscription, which brings together Grok, Grok Bot, Cursor, and X Premium in one plan. For the first time, one AI subscription supports the full path from thinking and creating to building and sharing. What’s included: Grok: chat, voice, and image and video generation. Grok Bot: AI teammates that get work done. Cursor: leading local and cloud coding agents. X: a checkmark, creator tools, and fewer ads, with higher tiers…
C
Where does the bug appear (feature/product)? Somewhere else… Describe the Bug In the Cursor web dashboard, changing the model variant of an automation doesn’t persist. I changed the automation’s model from “Grok 4.7 High Fast” to “Grok 4.7 Extra High” (the same happens with the other variant options like High, Extra High and Fast). When I leave the page and come back, the UI still shows the earlier option. The automation also runs with the default/earlier variant, Grok 4.7 High Fast, instead of…
C
Bug I sent one request — read and modify a single document. Cursor’s background workspace indexing / context gathering auto-routed to Grok 4.6 Medium, ignoring my BYOK (MiniMax) config, and consumed ~834k tokens in minutes. It appeared to traverse the entire codebase rather than just the one file I asked about. Steps to Reproduce Fresh install Cursor on new PC, configure BYOK (MiniMax) Open existing code project for the first time Send a single request: read and modify one document Check usage…
C
I recently saw how OpenAI was pushing and solving math problems and being bullish on Cursor Grok, I just gave the model a couple of different unsolved math problems and left it for a couple of hours. Out of the three I gave it, there had been progress on one. It managed to disprove a specific case of it. What really impressed me was that it did it in a couple of hours and only took a small part of the total usage(I am on the Pro Plan). The creating new math part is still under checking. If…
C
On the Free plan the model picker is locked to Cursor Grok 4.6 Medium. But Grok 4.7 costs the same as 4.6 per the pricing table Colin posted in the 4.7 announcement thread ($2 in / $0.50 cached / $6 out), and it’s the newer model, bigger base and better on long tasks. So why not lock free users to 4.7 instead? Free is where a lot of people get their first impression of Cursor, and right now that impression is the older model. If there’s a reason (capacity, SpaceXAI deal terms, rollout order,…
C
Is there a way to disable Grok as one of the options while using “Auto” settings? 2 posts - 2 participants Read full topic
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug With a Grok model selected, the usage UI shows the Other models meter. Grok 4.7, Grok 4.6, and Grok 4.5 are Cursor Models and should be shown against the Cursor Models pool. Steps to Reproduce Sign in on a plan that includes both pools (Pro, Pro Plus, or Ultra). Open the usage UI (editor settings or the Spending tab). Select a Grok model (Grok 4.7, Grok 4.6, or Grok 4.5) and send a request. Watch which meter the UI…
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug When creating a new chat with your last selected model set to Grok 4.5, the new chat auto selects Grok 4.7 instead of keeping it on Grok 4.5 Steps to Reproduce Start a new chat Select Grok 4.5 Write and submit a test prompt Start a new chat, selected model should show Grok 4.7 instead of Grok 4.5 Expected Behavior When I start a new chat the model I selected from the previous chat (Grok 4.5) auto changes to Grok 4.7. This…
C
Where does the bug appear (feature/product)? Cursor SDK Describe the Bug I call Cursor.models.list() and grok-4.7 advertises context values 256k and 500k. Its default variant is context 500k with reasoning_effort high and fast true. Agent.create accepts a selection with context 500k, but my run fails with: AI Model Not Found Invalid parameters for registry model: “grok-4.7”. The same high-fast selection with context 256k finishes. Passing only context 500k fails the same way. The SDK forwards…
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug When running an agent as Grok 4.7 Extra High, child workers cannot be spawned as Grok 4.7 because they are not in the “Live Task Allow List” From the model’s own diagnostic of this problem: This chat’s picker is already Grok 4.7, reasoning xhigh, Fast off. The CLI id for that variant is grok-4.7-xhigh. The live Task allowlist is still: composer-2.5-fast cursor-grok-4.6-high-fast gemini-3.6-flash-high gpt-5.6-sol-low Probe…
C
Main Announcement · Blog Now that Grok 4.7 is available, we’d love to hear how it’s working for you! Some things we’re especially curious about: How does it handle long, difficult tasks? Grok 4.7 was trained on a harder mix of problems weighted toward work that takes many hours, and it’s meant to verify its own work more carefully and manage longer context better than Grok 4.6. Does that hold up in your sessions? How is it on non-code knowledge work? This release was trained to natively…
C
I am having some major problems with the entire family of Grok products. I have a huge project and it is just destroying me with all of the issues that come out of no where. It’s a lot to go over but I am looking for a solution that can keep me from going bankrupt. Is there anyone here who can help me? I was wondering if I can have Fable correct the mistakes that Grok keeps making over and over? Would that just cause more problems? What Grok Bot says that Cursor is working on the problems is…
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug When I ask Gork‑4.6 to analyze and recognize a Figma design by pasting the Figma design link, an “Invalid Image” provider error is thrown. Request ID: 5d0d7a28-e848-4d48-a4cc-2bdc162aafea The error message says: Image error: The model provider could not read an image in this conversation. Animated GIFs, corrupted files, and images with extreme dimensions are rejected. Start a new chat without that image to continue. I only…
C
One example: I’m working on a browser extension that summarizes NYTimes ‘Opinion’ articles. It made a change, and I asked it to show me the new results. It answered by copying an earlier result that it found in the chat history. I’d appreciate any suggestions on how to ensure that stupid reasoning doesn’t take place. A transcript is available 8 posts - 4 participants Read full topic
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug I get this message in a lot of my requests and I don’t think I am using a sensitive prompt. Here is the error/notification : Request ID: 3d94c9ab-cfa5-4700-a42c-301c2c828a11 {“error”:“ERROR_OPENAI”,“details”:{“title”:“Request blocked”,“detail”:“We are unable to complete this request because it was blocked under the model provider’s usage guidelines. Try a less sensitive prompt.”,“isRe…
C
One of the most common way to save cost with LLMs right now is to have a smarter model do the planning and have implementation subagents using a less intelligent model. Given that the Projects feature heavily aligns with the concept of planner/worker model it’d be a really nice addition to be able to choose a planner and subagent model per project or for all projects. I love using Grok for execution but I think one of its biggest weaknesses is that it’s bad at high level stuff and can’t really…
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug Model picker on Cursor Automations displays only 4 models instead of all available models. By default it’s Grok 4.6 high-fast (but cannot be edited). Then it shows only 4 options (screenshot available). If other model is selected, Grok 4.6 high-fast option completely disappears. Model picker seems very different from regular model picker and must be fixed. Models across Cloud, IDE, CLI are a mess and must be unified.…
C
Feature request for product/service AI Models Describe the request I use Auto on Ultra. I don’t pick models by hand. Auto is supposed to be the cost-aware option, and Fast is the same Grok work at twice the token cost. I don’t need that speed for most of what I do. I need the monthly include to last. Fast is the default in Auto now. There is no setting to leave it off and only enable it when I want the extra speed. Cost, Balance, and Intelligence don’t give me that choice. Please add a setting…
C
Where does the bug appear (feature/product)? Cloud Agent (GitHub, Slack, Web, Linear) Describe the Bug On the mobile Cloud Agents interface, Cursor Grok 4.6 High is only available as “Cursor Grok 4.6 High Fast”. The normal/non-Fast version is completely missing from the model picker. Steps to Reproduce Open Cursor Cloud Agents on Android/mobile web. 2. Open the model picker. 3. Look for Grok 4.6 High. Only “Cursor Grok 4.6 High Fast” is shown; the non-Fast option is absent. Expected Behavior…
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug When creating automation from IDE/desktop app, I can’t choose normal (non-fast) grok 4.6. Only shows: cursor-grok-4.6-high-fast Cursor Grok 4.5 Fast Cursor Grok 4.5 High Fast Note I can toggle fast mode in WEB app, not IDE Steps to Reproduce IDE > Automations > New > model selector Operating System MacOS Version Information IDE: 3.18.9 (Universal) Does this stop you from using Cursor No - Cursor works, but with this issue…
C
Where does the bug appear (feature/product)? Cloud Agent (GitHub, Slack, Web, Linear) Describe the Bug When the chat/Cloud Agent picker is Cursor Grok 4.6 (High, Fast) and Settings → Models has only Grok 4.6 and Grok 4.5 enabled, Cloud Agent runs are still recorded as claude-4.5-sonnet (cloud icon on the Usage table). Claude is hidden by default and is not in my model list. I have never selected it. Explore Subagent Model is Default. This is not a one-off GPT call. I did use GPT once; that is…
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug cursor deleted models.py. problem timeline > cursor crashed and reloaded > cursor reload forces agent view > i close this (i don’t use it as I use IDE) > ssh session deleted from list > reloaded mulitple times > eventually opens > my remote server crashes and freezes > eventually I access remote via terminal and kill cursor process > cursor restarts > continue working and using grok > then conversation…
C
I see the API usage label while using cursor grok 4.6, is this counted towards the API usage pool rather than the model pool with cursor in house models? 4 posts - 2 participants Read full topic
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug Grok is enabled by default upon opening the IDE on mac. I’ve disabled it. I usually work in auto mode. For someone who despises X and Elon, it’s infuriating to know that our credits are filling his pockets. Feels like you’re getting a back hander from them. Please fix. Steps to Reproduce Open cursor, there it is. Operating System MacOS Version Information Version: 3.16.29 (Universal) VS Code Extension API: 1.128.0 Commit:…
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug The agent model picker does not keep the model I selected. On every sent message it resets to Grok 4.6 High Fast, even after I change it. This is extremely annoying: it overrides an explicit choice, can run the Fast variant I did not want, and forces me to re-select the model before every turn. This looks like the Grok 4.5 Fast-toggle / recommended-model override that Cursor staff already acknowledged as a known bug, now…
C
Main announcement · Blog Now that Grok 4.6 is available, we’d love to hear how it’s working for you. Some things we’re especially curious about: How does it handle long-running, multistep work? Grok 4.6 is built to stay with complex tasks across many steps, including more self-testing and verification before moving on. Does that match your experience? How is it on visual and interactive projects? We found it especially strong at turning a broad product idea into a working first version, with…
C
I am assuming this is whats happening but i’ve been only using Grok 4.5 and Cursor 2.5. i did not use any of these other models and seeing SONNET 5 and gemini 3.1 pro? I dont know whats going on, but thats ALOT of tokens spent for models i havent touched. Is there any way to fix this? I am using cursor-agent and i have the latest version. i dont even want to use cursor anymore because i dont want work done by sonnet5 or gemini 3.1 pro. 9 posts - 4 participants Read full topic
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug In the newest iteration of the agent window model selector, Grok 4.5 receives special treatment. When selected, the name of the model is not shown in the model selector under the text input buffer, just the effort level (e.g. High) which obscures the fact that Grok is selected. Other models are explicitly listed at the bottom of the agent window when selected. Steps to Reproduce Select Grok 4.5 (any effort level) as the…
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug I request that the agent spawns Cursor Grok 4.5 Medium when it is on Cursor Grok 4.5 High. It says that there is no such model available in the “allowed list”. Same goes for Cursor Grok 4.5 Low. So, when a Grok effort level is picked in the model selector, that is the only one it can use for agents. Doesn’t seem to be the case for other first party models like Composer 2.5 Steps to Reproduce Select “Cursor Grok 4.5 High”…