Model Identity Discrepancy and Performance Issues - Possible Billing Mismatch
Status Fixed / completed
Maintainer reply None cached
Activity 6 comments · opened Aug 4, 2025 · closed Aug 14, 2025
Bug Description
I'm experiencing significant issues with model consistency and performance that started today.
- The model is identifying itself as "Claude 3.5 Sonnet (October 2024 version)" (Max $200 user here)
- However, system prompts visible in the conversation reference "Opus 4" and model ID "claude-opus-4-20250514"
- I'm paying for Claude 4 access but appear to be receiving Claude Sonnet 3.5 responses, not even 3.7
- Performance has dramatically degraded - the model is making repetitive mistakes and behaving very differently compared to this morning
Evidence:
- Model self-identifies as Claude 3.5 Sonnet
- System messages reference Opus 4
- Significant behavior changes and error loops started suddenly
<img width="2138" height="224" alt="Image" src="https://github.com/user-attachments/assets/e448b8f7-0693-465f-96a3-b5ce41da3266" />
Request:
Please investigate whether there's a billing/access mismatch and ensure I'm receiving the correct model tier I'm paying for.
Environment Info
- Platform: darwin
- Terminal: vscode
- Version: 1.0.67
- Feedback ID: a9640454-45e5-4e3d-a766-4ca505802a13
Showing cached comments. Read the full discussion on GitHub ↗
4 Comments
Found 3 possible duplicate issues:
If your issue is a duplicate, please close it and 👍 the existing issue instead.
🤖 Generated with Claude Code
Yes i agree it has been like this for some time now i have noticed it's not as good they seem to be forcing us to use OPUS to burn credits knowing that sonnet is not doing the job, i used OPUS for my own planning not the plan mode they offer and then sonnet for coding, it worked a treat before with very few errors. NOW.. OMG it can't code anything hardly without having to do it 10 times and make a load of mistakes
it's really poor, would not be surprised if different regions get access to different AI chips on purpose too so there won't be any solid consistency to this issue
Things are looking good now after 1.0.70 update. Thank you Anthropic!
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.