It has always been like this. We actually know that the model performance has been mostly steady[0], but you cannot beat the 'evil companies secretly serve you worse models' memes. The meme value is too strong.

[0]: https://marginlab.ai/trackers/claude-code/

Are you using Opus? Sonnet remains as useful as it was while Opus efficacy and token burn rate has soured over the last 4 months.

When do you use it the most? I’ve noticed that it most often starts to degrade during 10-5 US East coast time. Late at night, I have the least amount of issues, but without fail, if I’m trying to do anything complex during the day, Claude gets loopy.

Same here. Works every time. Never ran into usage limits either.

Have you tried other harnesses, such as OpenCode?

Based on the experience of people using the $20 Claude Pro subscription and exhausting their quotas in a manner of minutes, the answer to your question is probably "less". (I would guess that the $100 plan would do the trick.)

Wow, I'm really surprised. I tried deepseek (their best model, through the official API). Its extremely cheap, but its clearly not as good at programming as Opus 4.7. It seems nowhere near as good at making high level design choices. Deepseek also seems to get stuck in whack-a-mole fixing loops much more than opus. I stopped it at one point, and asked opus to solve the problem it was trying to solve and it saw the solution immediately.

I was running deepseek through claude's code agent harness. Maybe it works better through a different tool?

interestingly I had the same experience, and weirdly it's in part because it is clearly less intelligent. It's more of a mechanistic tool just doing what I ask (but still very smart and very competent about it) and less trying to win a nobel prize with each answer. Turns out I actually like that.