C
Request ID: 8653f7e6-6b24-45ff-81a7-407fcd5b793e {“error”:“ERROR_OPENAI”,“details”:{“title”:“Request blocked by Anthropic”,“detail”:“We are unable to complete this request because it triggered restrictions on cyber content and was blocked under [Anthropic’s Usage Policy]( Real-time cyber safeguards on Claude Opus and Sonnet | Claude Help Center ). Try a less sensitive prompt.”,“isRetryable”:false,“additionalInfo”:{},“buttons”:[],“planChoices”:[],“analytics…
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug When connected via Remote-SSH to a Linux host whose default shell is zsh, Agent mode with Claude Opus 5.5 cannot run even trivial shell commands like echo hello. Both echo hello and /bin/echo hello fail with exit code 1 and: (eval):1: no matches found: [Console]::OutputEncoding The remote shell is zsh. [Console]::OutputEncoding is Windows PowerShell syntax, so it looks like the Agent Shell tool wrapper is injecting a…
C
With AI models constantly evolving, how do we actually distinguish one model from another? Is “newer = better” always valid when outputs are probabilistic and non-deterministic? Should we consider correctness, consistency, reasoning, context, speed, cost, and effort together? And how much does increasing effort affect the model’s actual performance? 1 post - 1 participant Read full topic
C
I’ve had this happen several times during this chat, Claude seems to answer in its thinking/working chunks and then that gets hidden before it’s shown to me. The resulting text is just a summary or if it shows a QA box, just the questions. Here’s the latest example where I asked it a point blank question to help understand its recommendation in response to it’s question, it thought for 18s and then gave me basically the same QA box back. Then it got perturbed when probed on it, claiming it…
C
Describe the Bug Choosing a model through @Cursor in slack doesn’t choose Fable 5.1 Steps to Reproduce In slack, @Cursor Use Fable 5.1 It replies using Fable 5 Expected Behavior It should choose Fable 5.1 Screenshots / Screen Recordings Operating System MacOS Version Information Version: 3.22.7 VS Code Extension API: 1.128.0 Commit: 37076c6c3f9e253c0fa2305197e45befd13a2260 Date: 2026-09-24T05:01:24.589Z Layout: Agent Window Build Type: Stable Release Track: Default Electron: 42.10.0 Chromium:…
C
Hello guys. I have been created a plan on “project”, the model for the project is opus 5.5. But even if I choose to build the plan with grok 4.7, it still spawning sub-agent in the project with opus 5.5 instead of the selected model. It burns out tons of my token for 3rd party model. I am not sure if this is an issue to be fixed or not. Thanks for helping 2 posts - 2 participants Read full topic
C
TL;DR: When a Claude safeguard trips in Cursor, the request is silently rerouted to Claude Opus 4.8, but the usage event is still recorded and billed as the model you picked (in my tests, claude-opus-5-5-max). There is no Opus 4.8 line item anywhere, which contradicts what Cursor staff have told users on this forum. All 3 of my triggered test requests behaved this way. Background and evidence are below. The billing breakdown is at the bottom. What Cursor told us When Fable 5 launched, a user…
C
Please share your experience on the Opus 5.5 model. Is it really that awesome? 11 posts - 8 participants Read full topic
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug Summary On us.anthropic.claude-opus-5-5, any chat containing a long multi-round agentic turn becomes permanently unusable. Every subsequent message fails with the same Anthropic 400. Short turns are unaffected — this only bites once a turn accumulates enough tool-call round-trips. The failing thinking block is always in the first assistant message, so it is re-sent on every later request and the thread can never recover.…
C
Where does the bug appear (feature/product)? Cursor CLI Describe the Bug In Cursor CLI, when we attempt to set the model to Opus 5.5 with a 1M context window, we can set the effort level, but upon returning to the main input screen, the effort level is set to MAX. This deviates from Cursor IDE, where we can set the model to Opus 5.5 with a 1M context window and the effort level remains set to what we intend for it to be. Steps to Reproduce In Cursor IDE: Set your model to Opus 5.5, with 1M…
C
Where does the bug appear (feature/product)? Cloud Agent (GitHub, Slack, Web, Linear) Describe the Bug In a Cloud Agent Project, worker agents created by the coordinator ignore my account’s Default Model (Preferences → Default Model = “Claude Opus 5.5 1M Medium Fast”). They always inherit the coordinator’s model (claude-opus-5-5-medium). The coordinator also can’t request the Fast or 1M variant explicitly, because those slugs aren’t in its allowed model list, even though the Fast variant is…
C
Describe the Bug My Opus 5.5 is stuck at 1m context window even though its set to 300k? It just burnt my token limit and wasted my plan… I only noticed when I checked the context break down… Steps to Reproduce Use Opus 5.5 - set to 300k Context. Expected Behavior The model does not use 1M context. Screenshots / Screen Recordings Operating System Windows 10/11 Version Information Version: 3.13.21 (system setup) VS Code Extension API: 1.128.0 Commit: 55434bd8062ece6fee083b82beed2aee42d253f0 Date:…
C
Opus 5.5 available at almost the same pricing as Grok 4.6 Fast Introducing Claude Opus 5.5 \ Anthropic Models & Pricing | Cursor Docs 2 posts - 2 participants Read full topic
C
Using e.g. pstack it is more or less must have to use 3rd party also. Grok 4.6 is fine but my experience is that Opus and Sol bring a lot the table. Solvable easily with money, but any other clever solutions? 7 posts - 4 participants Read full topic
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)? 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 After answering questions posed by Fable 5 the chat will often hang. Seems to be model-specific. Steps to Reproduce Start plan session with Fable as model Answer questions, hit enter/continue Observe (~75% occurrence rate for me) Expected Behavior Agent should resume with answers in conversation context. Screenshots / Screen Recordings Operating System MacOS Version Information Version: 3.17.8 (Universal) VS Code Extension…
C
Describe the Bug Why is my Claude Sonnet 5 model always is disable, everytime I restart the Cursor I need to re-enable it in the settings. Not only the Claude Sonnet 5 model always in disable, the other un-recommend officially model seems unstable as well, sometimes it’s appear sometimes it’s not Steps to Reproduce Restart the Cursor Expected Behavior Keep the enabled model always enabled Screenshots / Screen Recordings Operating System Windows 10/11 Version Information Version: 3.17.8 (user…
C
Where does the bug appear (feature/product)? Cursor IDE Describe the Bug Cursor not respecting model defined in subagent file when slash command used. I have an agent defined under Steps to Reproduce Create agent file in project under .cursor/agents/my-test-agent.md with the following configuration: --- name: my-test-agent description: Example agent model: 'claude-opus-5[effort=high]' color: cyan --- Open Cursor UI in workspace, type slash command, then submit it. /my-test-agent Expected…
C
Hey everyone, I recently set up AWS Bedrock in Cursor. Fable 5 didn’t show up in the default dropdown list from Bedrock, so I it manually at the bottom as a custom model using the exact cross-region string: us.anthropic.claude-fable-5 The connection works fine, but context window is just 200K. Has anyone else experienced this? 3 posts - 2 participants Read full topic
C
Feature request for product/service AI Models Describe the request I already pay for Claude Max. Why can’t I use it properly inside Cursor? Cursor has the best visual coding experience I have seen so far. When an agent edits code in Cursor, I can actually follow what is happening: file changes, diffs, editor context, applied edits, review flow. It feels natural for real development work. But here is the problem: Many of us already pay for Claude Max because Claude is one of the strongest coding…
C
Hi, I’ve set up Cursor ACP in IntelliJ IDEA and got all the models shown in the window. It works fine with auto. It told me to add a payment method when I’m using Opus 4.6. I’ve checked my limit and payment settings, and it works fine in Cursor GUI. My limit is currently at $100 / $200. 4 posts - 4 participants Read full topic
C
Where does the bug appear (feature/product)? Background Agent (GitHub, Slack, Web, Linear) Describe the Bug When a user configures their own Anthropic API key (Cursor Settings > Models > Anthropic API Key), the parent agent correctly routes requests through the user’s key. However, subagents spawned via the Task tool are not routed through the user’s own key — they are billed against the Cursor plan instead. This causes two distinct failure scenarios depending on the user’s plan state: Scenario…