[BUG] Sonnet 4 is completely unsuable because of 429s
Status Fixed / completed
Maintainer reply ✓ Yes — bcherny
Activity 7 comments · opened May 30, 2025 · closed May 30, 2025
💡 Likely answer: A maintainer (bcherny, collaborator)
responded on this thread — see the highlighted reply below.
Environment
- Platform (select one):
- [ ] Anthropic API
- [x] AWS Bedrock
- [ ] Google Vertex AI
- [ ] Other: <!-- specify -->
- Claude CLI version: 1.0.6 (Claude Code)
- Operating System: Linux (Fedora 41, 6.13.5-200.fc41.x86_64 (64-bit))
- Terminal: warp.dev
Bug Description
Constant stream of 429s for sonnet 4 model when using claude code API Error (429 Too many tokens, please wait before trying again.) · Retrying in 1 seconds… (attempt 1/10). I am completely unable to use sonnet 4.
It is not AWS service quota limitation. I tried with extremely low message rate.
This happens for arbitrarily simple and small messages. The cwd had very little content.
It also happens on sonnet 3.7, but its more rare for 3.7 and 3.7 is usable.
Steps to Reproduce
- Fresh install of claude code
- Set
AWS_DEFAULT_REGION,AWS_PROFILEand other AWS authentication environment variables - Set model to sonnet 4
- Type a message
- I get stream of constant 429s
Expected Behavior
429s should happen only when I hit AWS service quotas.
Actual Behavior
I get stream of 429s
Additional Context
Using cross-region inference profile, region: us-west-2,
7 Comments
same in us-east-1
Yeah, this is already known:
https://github.com/anthropics/claude-code/issues/1347
https://github.com/anthropics/claude-code/issues/1293
Hi! Please request a rate limit increase in your Bedrock Console. Docs here: https://docs.aws.amazon.com/bedrock/latest/userguide/quotas.html#quotas-increase.
It is not a service quota limitation. It happens even at arbitrarily low rates. I am no where close to hitting quota.
I had the same issue when trying Bedrock for Claude Code yesterday. I can even send requests using the SAME IAM credentials to InvokeModel at the same time while the errors are happening in Claude Code with other tools and they work completely fine. I had a similar issue while trying Bedrock with Aider and was only able to use it there with the Converse API, not with the default InvokeModel API. Something must be up with how the Bedrock API is called or some weird issue on the Bedrock platform.
Just want to +1 this. Not sure how to get this on someone's radar at Anthropic + AWS, but it's making Claude Code unusable in enterprise environments that mandate AWS Bedrock usage.
This issue has been automatically locked since it was closed and has not had any activity for 7 days. If you're experiencing a similar issue, please file a new issue and reference this one if it's relevant.