[BUG] After the release of Opus 4.1, the entire Claude 4 lineup stopped writing working code.
Status Closed — not planned
Maintainer reply None cached
Activity 11 comments · opened Aug 11, 2025 · closed Jan 8, 2026
This happened immediately after the release of Opus 4.1. All code that the AI outputs is now practically non-functional. There are many errors, the AI doesn't follow prompts, makes fundamentally wrong decisions on its own, and breaks things that were working before.
This affected the entire Claude 4 lineup. Opus 4 and Sonnet 4 now also write non-working code, even though before the release of Opus 4.1 they wrote complex code with virtually no errors.
11 Comments
Found 3 possible duplicate issues:
This issue will be automatically closed as a duplicate in 3 days.
🤖 Generated with Claude Code
this has definitely been my experience this week 5 days claude opus 4.1 max 20
I simply believe that the problem is context dropout for lack of a more accurate term. It wil lose tons context, even in projects and you ahve to continuously go over the context and ad to it. It will drop large sections of functional code, it will get intop refactoring loops that lead nowhere. All-in-all a very difficult programming experience, likely a very buggy AI.
Has anything changed since then?
In effect I resolved my issue and Claude is pumping out much better code.
+1 no new model is behaving how it did prior to these releases. I've actually noticed this in the Claude web app. When an artifact already exists and claude tries to edit it, more times than not it just breaks the artifact by inserting code in the wrong place, or just outputting the same artifact multiple times. I've had to resort to asking Claude to "tell me what to change and where to change it" rather than Claude doing it itself.
Same for me with Pro subscription.
Sonnet been behaving like Haiku over last several weeks.
Which instantly felt like model becoming dumber, missing existing context and hallucinating non-existent.
Was going to upgrade to Max, but now considering moving to Cerebras Code Max with Qwen3-Coder instead.
My guesstimate is that Anthropic is downgrading live inference with "optimized" quantized precision during peak hours (e.g. FP8/16 -> FP4), either tuning context memory utilization with some sort of lossy flash attention.
I can understand that fix-price subscriptions does not fully cover massive consumption and are largely subsidized.
Latency degradation and 5hr block time limits are acceptable.
But silent quality degradation creates UNRELIABLE REPUTATION.
It's a terrible feel of being gaslighted instead of transparent warnings for quality degradation or model downgrade.
@gudzenkov +1.
Sometimes I wonder if I’m working in ‘auto’ mode, since the output has lost consistency regardless. Even if I choose and use an advanced model for a long time. At some point of those long work sessions I end up feeling paranoid.
Yeah, I noticed that yesterday too — I was giving it simple tasks, but it kept doing strange things. It even added corrections to problems that had already been solved just a couple of minutes earlier. I used Pro subscription.
This issue has been inactive for 30 days. If the issue is still occurring, please comment to let us know. Otherwise, this issue will be automatically closed in 30 days for housekeeping purposes.
This issue has been automatically closed due to 60 days of inactivity. If you're still experiencing this issue, please open a new issue with updated information.
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.