Planning next moves stuck

Support Bug Reports
107

Yes for last 2 days, I am also seeing same behavior. No matter what model we use, it is so slow that it feels like I should code it myself. Please fix this urgently.

3 Likes
108

same

1 Like
109

even after using http/2.0 or using 1.1 i dont thing there are many different. even sometime, using 1.1 not change much

1 Like
110

I don’t know, but if I change to http/1.1, the Agent takes a long time to load without any response. http/2 doesn’t have this error.

I’ve had this error since upgrading to the new Cursor version. Last week everything was fine with http/1.1

2 Likes
111

same for me since yesterday.

1 Like
112

Same, I even got many “try again“. I hope I can resume the conversation.

1 Like
113

Using GPT-5.1

114

Hey everyone, thanks for the report.

Please try:

  • HTTP/2 setting: open Settings, find “HTTP/2”, enable “Disable HTTP/2”, restart Cursor, and test. If it didn’t help, turn it back off and test again
  • Create a brand-new chat in an empty workspace and test 1–2 models

And please share:

  • 2–3 Request IDs from slow/error messages (Chat message context > Copy Request ID)
  • A screenshot from Settings > Network > Run Diagnostics

Let me know if the steps above helped. If not, I’ll collect everything and escalate to the team.

1 Like
115

saved my day

116

same

117

All models are same too slow！Why is it so slow?

118

The previous version seemed so much faster. Why does it get slower with each upgrade?

119

All agents are respo`nding very slow, is there any outage?

120

same since today

121

yes, whom to reach out?

122

Where does the bug appear (feature/product)?

Cursor CLI

Describe the Bug

Cursor’s agent mode is extremely slow, it gets stuck in planning mode for a long time after every request.

Steps to Reproduce

Open a chat in agent mode, submit the request => gets stuck in “Planning next moves”;
I tried to use both “Auto” and specific models available for me, the result is the same

Expected Behavior

Much faster responses

Operating System

MacOS

Current Cursor Version (Menu → About Cursor → Copy)

Version: 2.1.50
VSCode Version: 1.105.1
Commit: 56f0a83df8e9eb48585fcc4858a9440db4cc7770
Date: 2025-12-06T23:39:52.834Z (2 days ago)
Electron: 37.7.0
Chromium: 138.0.7204.251
Node.js: 22.20.0
V8: 13.8.258.32-electron.0
OS: Darwin arm64 25.1.0

For AI issues: which model did you use?

I tried “Auto”, and other various models, like Sonnet 4.5, GPT-5.1 Codex, Gemini 3 Pro

For AI issues: add Request ID with privacy disabled

Request ID: af129e80-8819-4b90-998a-ed48212a06ec

Does this stop you from using Cursor

Sometimes - I can sometimes use Cursor

123

My HTTP2 was already off. Turning it back on and restarting fixed this issue, especially for Composer

124

same。

Since December 8th, I have been experiencing extreme lag and instability. The specific symptoms are as follows:

  1. Message Failure & Errors: Frequently, messages sent to the Agent fail to deliver. I have encountered the “extension host error” multiple times. Reloading the window does not resolve the issue.

  2. Severe Latency: The Agent has become incredibly slow during task execution. It frequently hangs on steps like “planning,” “reading files,” or “executing code.” Tasks that previously took 3-4 minutes now require over 30 minutes to complete.

  3. Execution Interruptions: The Agent often interrupts unexpectedly during tasks, forcing me to manually retry multiple times.

Troubleshooting Steps Taken: I have already attempted the following with no success:

  • Restarted Cursor multiple times.

  • Restarted my computer.

  • Completely uninstalled Cursor (including clearing cache and configuration files) and reinstalled it.

Additional Context: A colleague of mine (also a Cursor Ultra user) has been experiencing similar issues since December 8th. They are currently on Cursor Version 2.1.48. This suggests the issue may be widespread or server-side.

Some Diagnostic Results:

  1. Network: I have run the diagnostics multiple times (including just now). The results show the network is Normal.

  2. Models: I tested across GPT, Claude, and Gemini series models. The latency and failure issues persist regardless of the model selected.

  3. Environment: I initially suspected my remote server connection, so I tested on a purely local project as well. The issue remains the same.

My System Information:

  • Version: 2.1.50

  • VSCode Version: 1.105.1

  • Commit: 56f0a83df8e9eb48585fcc4858a9440db4cc7770

  • Date: 2025-12-06T23:39:52.834Z

  • Electron: 37.7.0

  • Chromium: 138.0.7204.251

  • Node.js: 22.20.0

  • V8: 13.8.258.32-electron.0

  • OS: Darwin arm64 24.1.0

2 Likes
125

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

All models are responding very slow from today.

Steps to Reproduce

Ask anything, the agent stuck in planning mode.

Expected Behavior

Should work fast

Operating System

MacOS

Current Cursor Version (Menu → About Cursor → Copy)

Version: 2.1.50 (Universal)
VSCode Version: 1.105.1
Commit: 56f0a83df8e9eb48585fcc4858a9440db4cc7770
Date: 2025-12-06T23:39:52.834Z
Electron: 37.7.0
Chromium: 138.0.7204.251
Node.js: 22.20.0
V8: 13.8.258.32-electron.0
OS: Darwin arm64 25.1.0

For AI issues: which model did you use?

Auto, Opus, and others

Does this stop you from using Cursor

Yes - Cursor is unusable

126

I have done all the checks mentioned in this thread, and everything is set, i.e, I disabled HTTP2, Network diagnostic doesn’t show any issues, and I started Cursor with disabled extensions, however this didn’t seem to have any affect on the behavior, the issue remains.