›_ Claude Code Issues
Claude Code Issues / #28760

[Bug] Rate Limit Error persists after increasing plan limits and adding funds and Max plan only at 33% used

Open 💬 24 comments Opened Feb 25, 2026 by Jcashiola

Bug Description
I have the max plan. Claude code used up $50 worth of overflow. My plans usage is a 1% session and 33% daily. I added funds and increased the limits in hopes of clearing the Rate limit Errod. but no help. WHat do I do?

Environment Info

  • Platform: win32
  • Terminal: windows-terminal
  • Version: 2.1.58
  • Feedback ID: d558c6e4-d958-44ea-bcbb-2b8611f4e8e4

Errors

[{"error":"Error: 429 {\"type\":\"error\",\"error\":{\"type\":\"rate_limit_error\",\"message\":\"Extra usage is required for long context requests.\"},\"request_id\":\"req_011CYVYGFpSu4dLP39BBWbKh\"}\n    at W4.generate (file:///C:/Users/jcash/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:329:28354)\n    at qh.makeStatusError (file:///C:/Users/jcash/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:368:2195)\n    at qh.makeRequest (file:///C:/Users/jcash/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:368:5420)\n    at process.processTicksAndRejections (node:internal/process/task_queues:103:5)","timestamp":"2026-02-25T22:00:48.603Z"}]

View original on GitHub ↗

24 Comments

github-actions[bot] · 4 months ago

Found 3 possible duplicate issues:

  1. https://github.com/anthropics/claude-code/issues/27603
  2. https://github.com/anthropics/claude-code/issues/27179
  3. https://github.com/anthropics/claude-code/issues/28096

This issue will be automatically closed as a duplicate in 3 days.

  • If your issue is a duplicate, please close it and 👍 the existing issue instead
  • To prevent auto-closure, add a comment or 👎 this comment

🤖 Generated with Claude Code

andrewleejenkins · 4 months ago

Having the same issue. Unable to use this without extra usage and I'm coming back after a 10 hour break of no use on Max. Should have plenty of headway available. started about 12 hours ago.

Jcashiola · 4 months ago

I added $5 to the account. The rate limit error went away. Then monitored the usage. I still have the $5. I've been working ever since no issues. You could try $1. My guess is the $5 is gone.

Apollo1113 · 4 months ago

I am also having this issue on sonnet 4.6 claude desktop cowork latest version. does not work even if switching to opus 4.6

Jcashiola · 4 months ago

I spoke too soon. Watch out. I was at the end of a CC session. The rate limit error came back with a vengeance. It ate my $5. I put in another $5. Gone in a minute. Put $10 just to try to finish. Gone in a little over 2 minutes. Put in another $10. Same thing, Gone in a little over 2 minutes. A single agent was running. Turned it off. Right at $5 a minute. No response from support.

zyqxd · 4 months ago

Same, switched from PAYG to Max 5x. Getting 429 with $60+ left over and no sessions used. MacOS

kevanghobadi · 4 months ago

Same issue here on max plan with MacOS

clementgineste · 4 months ago

Same issue, happened few times over last days. This is due to a potentially corrupt session. If you open a new one, another folder, you'll be able to query models again.
If you succeed to /compact you'll gain back access. Usually not working when changing to Sonnet but always worked with Haiku.

Exemple, right now:

❯ ls                                                                                                                                                                     
  ⎿  Async hook UserPromptSubmit completed                                                                                                                               
  ⎿  Async hook Stop completed                                                                                                                                           
  ⎿  API Error: Rate limit reached                                                                                                                                       
                                                                                                                                                                         
❯ /context                                                                                                                                                               
  ⎿  Context Usage                                                                                                                                                       
     ⛁ ⛀ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁   claude-opus-4-6 · 76k/200k tokens (38%)                                                                                                       
     ⛀ ⛁ ⛁ ⛀ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁                                                                                                                                                 
     ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁   Estimated usage by category                                                                                                                   
     ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁   ⛁ System prompt: 3.4k tokens (1.7%)                                                                                
     ⛁ ⛁ ⛀ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶   ⛁ System tools: 17.1k tokens (8.6%)                                                                                                           
     ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶   ⛁ Memory files: 5k tokens (2.5%)                                                                                                              
     ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶   ⛁ Skills: 457 tokens (0.2%)                                                                                                                   
     ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶   ⛁ Messages: 57.4k tokens (28.7%)
     ⛶ ⛶ ⛶ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝   ⛶ Free space: 84k (41.8%)
     ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝   ⛝ Autocompact buffer: 33k tokens (16.5%)

     Memory files · /memory
     └ CLAUDE.md: 3.6k tokens
     └ ~/.claude/projects/-work-anon-career/memory/MEMORY.md: 1.4k tokens

     Skills · /skills

     Project
     └ analyze-offer: 16 tokens
     └ generate-cv: 16 tokens
     └ cover-letter: 16 tokens

     User
     └ peon-ping-config: 67 tokens
     └ peon-ping-use: 64 tokens
     └ peon-ping-toggle: 62 tokens
     └ peon-ping-log: 52 tokens

❯ ls                                                                                                                                                                     
  ⎿  Async hook UserPromptSubmit completed
  ⎿  Async hook Stop completed                                                                                                                                           
  ⎿  API Error: Rate limit reached
                                                                                                                                                                         
❯ /compact                                                                                                                                                               
  ⎿  Error: Error during compaction: Error: API Error: Rate limit reached

Changing to Sonnet:

❯ /model sonnet                                                                                                                                                          
  ⎿  Set model to sonnet (claude-sonnet-4-6)                                                                                                                             
                                                                                                                                                                         
❯ /compact                                                                                                                                                               
  ⎿  Error: Error during compaction: Error: API Error: Rate limit reached

And finally Haiku:

❯ /compact 
  ⎿  Compacted (ctrl+o to see full summary)
  ⎿  Read output/cv-alan-platform-engineer.md (96 lines)
  ⎿  Read output/cv-sopra-devops-telco.md (98 lines)
  ⎿  Read data/profile.yaml (47 lines)
  ⎿  Read data/projects.yaml (121 lines)
  ⎿  Read data/experiences.yaml (236 lines)
  ⎿  Plan file referenced (~/.claude/plans/glowing-yawning-canyon.md)

❯ ls                                                                                                                                                                     
  ⎿  Async hook UserPromptSubmit completed
  ⎿  Async hook Stop completed                                                                                                                                           
                                         
● Looking at the context, I see you've completed several tasks in the previous session:                                                                                  
   
  1. ✅ X hiring team message — Written and iterated to 3 short paragraphs (approved)                                                                         
  2. ✅ Analytics tracking for cv repo — GitHub issue #36 opened with GoatCounter recommendation
  3. ✅ Y Platform Engineer CV — Generated and saved to output/cv-alan-platform-engineer.md (37.4KB)                                                                  
                                                            
  All recent requests have been completed. What would you like to work on next?

❯ /context                                                                                                                                                               
  ⎿  Context Usage
     ⛁ ⛀ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁   claude-haiku-4-5-20251001 · 42k/200k tokens (21%)                                                                                             
     ⛀ ⛁ ⛁ ⛀ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁                                                                 
     ⛁ ⛁ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶   Estimated usage by category                                                                                                                   
     ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶   ⛁ System prompt: 3.4k tokens (1.7%)                                                                                
     ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶   ⛁ System tools: 17.1k tokens (8.6%)                                                                                                           
     ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶   ⛁ Memory files: 5k tokens (2.5%)                                                                                                            
     ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶   ⛁ Skills: 457 tokens (0.2%)                                                                                                                   
     ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶   ⛁ Messages: 16.9k tokens (8.5%)
     ⛶ ⛶ ⛶ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝   ⛶ Free space: 124k (62.0%)
     ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝   ⛝ Autocompact buffer: 33k tokens (16.5%)

     Memory files · /memory
     └ CLAUDE.md: 3.6k tokens
     └ ~/.claude/projects/-work-anon-career/memory/MEMORY.md: 1.4k tokens

     Skills · /skills

     Project
     └ analyze-offer: 16 tokens
     └ generate-cv: 16 tokens
     └ cover-letter: 16 tokens

     User
     └ peon-ping-config: 67 tokens
     └ peon-ping-use: 64 tokens
     └ peon-ping-toggle: 62 tokens
     └ peon-ping-log: 52 tokens

❯ What model are you ?                                                                                                                                                   
  ⎿  Async hook UserPromptSubmit completed
  ⎿  Async hook Stop completed                                                                                                                                           
                                         
● I'm running on Claude Haiku 4.5 (claude-haiku-4-5-20251001).                                                                                                           
   
  You're at 21% context usage (42k/200k tokens), with about 124k tokens of free space remaining. You can easily continue working without hitting context limits.         
                                                            
❯ /model opus                                                                                                                                                            
  ⎿  Set model to opus (claude-opus-4-6)                                                                                                                                 
                                                                                                                                                                         
❯ What model are you ?                                                                                                                                                   
  ⎿  Async hook UserPromptSubmit completed                                                                                                                               
  ⎿  Async hook Stop completed                                                                                                                                           
                                                                                                                                                                         
● I'm running on Claude Opus 4.6 (claude-opus-4-6). You just switched from Haiku 4.5.
Jcashiola · 4 months ago

It went away when I turned off the 1m context window.

On Mon, Mar 2, 2026 at 4:46 AM Clément Gineste @.***>
wrote:

clementgineste left a comment (anthropics/claude-code#28760) <https://github.com/anthropics/claude-code/issues/28760#issuecomment-3983879018> Same issue, happened few times over last days. This is due to a potentially corrupt session. If you open a new one, another folder, you'll be able to query models again. If you succeed to /compact you'll gain back access. Usually not working when changing to Sonnet but always worked with Haiku. Exemple, right now: ❯ ls ⎿ Async hook UserPromptSubmit completed ⎿ Async hook Stop completed ⎿ API Error: Rate limit reached ❯ /context ⎿ Context Usage ⛁ ⛀ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ claude-opus-4-6 · 76k/200k tokens (38%) ⛀ ⛁ ⛁ ⛀ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ Estimated usage by category ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ System prompt: 3.4k tokens (1.7%) ⛁ ⛁ ⛀ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛁ System tools: 17.1k tokens (8.6%) ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛁ Memory files: 5k tokens (2.5%) ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛁ Skills: 457 tokens (0.2%) ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛁ Messages: 57.4k tokens (28.7%) ⛶ ⛶ ⛶ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛶ Free space: 84k (41.8%) ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ Autocompact buffer: 33k tokens (16.5%) Memory files · /memory └ CLAUDE.md: 3.6k tokens └ ~/.claude/projects/-work-anon-career/memory/MEMORY.md: 1.4k tokens Skills · /skills Project └ analyze-offer: 16 tokens └ generate-cv: 16 tokens └ cover-letter: 16 tokens User └ peon-ping-config: 67 tokens └ peon-ping-use: 64 tokens └ peon-ping-toggle: 62 tokens └ peon-ping-log: 52 tokens ❯ ls ⎿ Async hook UserPromptSubmit completed ⎿ Async hook Stop completed ⎿ API Error: Rate limit reached ❯ /compact ⎿ Error: Error during compaction: Error: API Error: Rate limit reached Changing to Sonnet: ❯ /model sonnet ⎿ Set model to sonnet (claude-sonnet-4-6) ❯ /compact ⎿ Error: Error during compaction: Error: API Error: Rate limit reached And finally Haiku: ❯ /compact ⎿ Compacted (ctrl+o to see full summary) ⎿ Read output/cv-alan-platform-engineer.md (96 lines) ⎿ Read output/cv-sopra-devops-telco.md (98 lines) ⎿ Read data/profile.yaml (47 lines) ⎿ Read data/projects.yaml (121 lines) ⎿ Read data/experiences.yaml (236 lines) ⎿ Plan file referenced (~/.claude/plans/glowing-yawning-canyon.md) ❯ ls ⎿ Async hook UserPromptSubmit completed ⎿ Async hook Stop completed ● Looking at the context, I see you've completed several tasks in the previous session: 1. ✅ X hiring team message — Written and iterated to 3 short paragraphs (approved) 2. ✅ Analytics tracking for cv repo — GitHub issue #36 opened with GoatCounter recommendation 3. ✅ Y Platform Engineer CV — Generated and saved to output/cv-alan-platform-engineer.md (37.4KB) All recent requests have been completed. What would you like to work on next? ❯ /context ⎿ Context Usage ⛁ ⛀ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ claude-haiku-4-5-20251001 · 42k/200k tokens (21%) ⛀ ⛁ ⛁ ⛀ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛁ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ Estimated usage by category ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛁ System prompt: 3.4k tokens (1.7%) ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛁ System tools: 17.1k tokens (8.6%) ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛁ Memory files: 5k tokens (2.5%) ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛁ Skills: 457 tokens (0.2%) ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛶ ⛁ Messages: 16.9k tokens (8.5%) ⛶ ⛶ ⛶ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛶ Free space: 124k (62.0%) ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ ⛝ Autocompact buffer: 33k tokens (16.5%) Memory files · /memory └ CLAUDE.md: 3.6k tokens └ ~/.claude/projects/-work-anon-career/memory/MEMORY.md: 1.4k tokens Skills · /skills Project └ analyze-offer: 16 tokens └ generate-cv: 16 tokens └ cover-letter: 16 tokens User └ peon-ping-config: 67 tokens └ peon-ping-use: 64 tokens └ peon-ping-toggle: 62 tokens └ peon-ping-log: 52 tokens ❯ What model are you ? ⎿ Async hook UserPromptSubmit completed ⎿ Async hook Stop completed ● I'm running on Claude Haiku 4.5 (claude-haiku-4-5-20251001). You're at 21% context usage (42k/200k tokens), with about 124k tokens of free space remaining. You can easily continue working without hitting context limits. ❯ /model opus ⎿ Set model to opus (claude-opus-4-6) ❯ What model are you ? ⎿ Async hook UserPromptSubmit completed ⎿ Async hook Stop completed ● I'm running on Claude Opus 4.6 (claude-opus-4-6). You just switched from Haiku 4.5. — Reply to this email directly, view it on GitHub <https://github.com/anthropics/claude-code/issues/28760#issuecomment-3983879018>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AW5CIGEAYYERMZWUTSWDPFT4OVYAVAVCNFSM6AAAAACV72DM7SVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTSOBTHA3TSMBRHA> . You are receiving this because you authored the thread.Message ID: @.***>
oliverstreissi · 4 months ago

Had the same issue because of the 1M context window, although I never set that. Could only turn it off by setting CLAUDE_CODE_DISABLE_1M_CONTEXT=1. Now it's working again.

clementgineste · 4 months ago
Had the same issue because of the 1M context window, although I never set that. Could only turn it off by setting CLAUDE_CODE_DISABLE_1M_CONTEXT=1. Now it's working again.

Just worked for me. And directly it compacted our conversation (i was at 80%).

edit: arf, came back after 1 question ...
edit2: remove disable_1m and it came back to live ,wtf ?

clementgineste · 4 months ago

Again and again, so annoying

andrewleejenkins · 4 months ago

Yeah for anybody wondering change the /model to not use the 1M context window. its a bug where you can overshoot your limit and get stuck on extra usage.

clementgineste · 4 months ago

Deleted my ~/.claude/projects and it's working, pretty sure something is wrong with old .jsonl in i don't know with case and boom

clementgineste · 4 months ago

Ok came back.

you got hooks / plugins ?

just peon-ping here

aplathrop-qb · 4 months ago

Same issue for me. As others have said I had hit a limit on 5X max for my session. Upgraded to 20 X and still hit the API rate limit error. Spun around on it for 24 hours logging in to the API version / logging out, etc. try lots of bugging in the leading files but none of that worked. I then switched to the non-1,000,000 context window and that just started working for me.

Mobz22 · 4 months ago

Currently having this issues on 20x Plan, my weekly usage is 10%, by my extra usage went to 100%, now getting API Error: Rate limit reached. Also I have no issues using Claude Desktop code

bottyBotz · 4 months ago

Seeing this issue as well on my Team plan.
Current Session - 0% used
Weekly limits all models - 69% used.

Issue only presents itself on Claude Cowork at this time.

codenics · 4 months ago

<img width="442" height="65" alt="Image" src="https://github.com/user-attachments/assets/a5fd9f7e-9764-4288-811a-404233e5f05d" />

1.Turned _auto-reload_ on in _Settings -> Usage_.
2.Relaunched _Claude Code_ and it worked!

jedijamez567 · 4 months ago

This is how I fixed this:

First, backup ~/.claude.json --> ~/.claude.json.bak

Next, claude logout

next, rm -f ~/.claude.json

next, claude login

Finally, restore original claude.json: mv ~/.claude.json.bak ~/.claude.json

It should work after this.

tonygaglio · 4 months ago

Having the same problem, to fix, switch to another model that isn't the 1M context one.

iRonin · 4 months ago

tried many things but finally switching between Sonnet and Default worked for me

dvmer · 4 months ago

Same problem. I had to turn on the toggle for the extra usage for it to work.

byteolu · 3 months ago

If you’re using a new account, go into Claude console -> Organization setting -> Billing and make sure you have a credit balance.