[BUG] Claude 4 (or since its release) is BURNING USAGE LIMITS x10 TIMES FASTER

Yes, I saw it too. I updated my plan up.

While the default for Claude Max users is to use the "best available" model (Opus until you reach rate limits, then Sonnet), you can choose to pin a specific model.

• claude --model flag will be used for the duration of the session.
• Within Claude Code, /model will switch the model for the session and save your preference for future sessions.

Examples:

# Always use the latest sonnet, currently claude-sonnet-4-20250514
claude --model sonnet
# Always use the latest opus, currently claude-opus-4-20250514
claude --model opus
# Always use a specific model. Provide the full model name
# https://docs.anthropic.com/en/docs/about-claude/models/all-models#model-names
claude --model claude-3-7-sonnet-20250219

Or once you're in Claude Code

> /model (to open the selection menu)
> /model sonnet (to stick to the latest sonnet, currently claude-sonnet-4-20250514)
> /model claude-3-7-sonnet-20250219 (to use a specific model indefinitely)

We have updated the documentation about model configuration in https://docs.anthropic.com/en/docs/claude-code/bedrock-vertex-proxies#model-configuration to reflect these options.

This is not being a problem with models Settings for two reasons:

1) I never switched from models and always used the default one that I guess is Opus and should be working exactly the same way that before claude's 4 release.

2) to continue working I had to come back to normal claude's chate page and surprise, I added the file which I am working with 3000 code lines approx and surprise! I reached the limit in 30 minutes! 4 interacions in a total of generating 12.000 code lines on a 35Kb text file.

CONCLUSION: Now Claude has on its MAX plan less usage than on basic $14 plan before the release. WTF?
I was working like charm even with normal Claude chat with more than 30 files on context and +80.000 lines of code for HOURS an I BURNT IT IN 30 MINUTES with 4 requests !!!

THAT CAN ONLY BE A SCAM STRAGEGY RASING DOWN THE USAGE LIMITS!

Or... I stil want to beleave, they didn't considered that new models are consumming as hell and they have to urgently rise up all limits to be on same level than claude 3.7.

@dicksontsai, just after reach rate limits, and switch to Sonet, Claude Code is not working. So I must only up the price plan.

@EddiePiazza, all will be okay. Just they need some time, to see some strange statistics, so they will fix it.

Hope so, I was just right now being blocked again reaching limit after 2 hours coding 1 single file having to wait 3 hours till continue. It's impossible to work...

@dicksontsai Just to let you know, I totally disabled al 6pm all logs and memory fom claude to prevent fill claude with long contexts. after 2 hour working Bie Bie, reached the max usage again. That is just crazy becoming from such company, I really hope this is a Bug of Wrong metrics measurement because if it's a new business strategy, the answer is just DOESN'T WORTH ANYMORE 1440€/YEAR

I'm a little confused on why this issue was closed @dicksontsai ? It Anthropic ignoring this concern? If someone is using the product the way they were using it before and are achieve significantly less success than prior to upgrading to the 4.0 models, this seems like a serious concern that needs addressing on Anthropic's end, or you certainly should NOT default to a model that exhausts even the basic MAX plan within 2 hours barely without writing a large amount of code. If this is expected behavior then you should be providing your users with "upgrade guides" on how to continue to use the same product (Claude Code) without having to "figure out" how to completely shift their processes when new models come out.

well, @dicksontsai in fact @dgarson is on the right way. This issue should not be closed as more users will complain about it. I was forced to use Sonnet to get hugest limits but that's not acceptable as Claude 4 is DOWNGRADING our expensive MAX plans forcing all people to use always Sonnet as it's impossible to use Opus. The only reason now to use Opus it to intend to work for not longer than 2 to 3 hours. Since using Sonnet I am able to work for all day but I am getting a lot of errors and wasting time in comparison with Opus. I feel Scammed....

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.