Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Pro Max 5x quota exhausted in 1.5 hours despite moderate usage

github.com

Post reply on HN
Print view View original

291–300 of 695 posts

Re: Pro Max 5x quota exhausted in 1.5 hours despite moderate usage

#291

Post by aaronblohowiak »

bcherny wrote: post #257
mvkel wrote: post #227

Earlier quoted context omitted.

Why did this become an issue seemingly overnight when 1M context has been available for a while, and I assume prompt caching behavior hasn't changed? EDIT: prompt caching behavior -did- change! 1hr -> 5min on March 6th. I'm not sure how starting a fresh session fixes it, as it's just rebuilding everything. Why even make this available? It feels like the rules changed and the attitude from Anth is "aw I'm sorry you di…

> 1hr -> 5min on March 6th This is not accurate. The main agent typically uses a 1h cache (except for API customers, which can enable 1h but it is not on by default because it costs more). Sub-agents typically use a 5m cache.

So if I run a test suite or compile my rust program in a sub agent I’m going to get cache misses? Boo.

Re: Pro Max 5x quota exhausted in 1.5 hours despite moderate usage

#293

Post by bcherny »

reenorap wrote: post #290
bcherny wrote: post #210

Hey all, Boris from the Claude Code team here. We've been investigating these reports, and a few of the top issues we've found are: 1. Prompt cache misses when using 1M token context window are expensive. Since Claude Code uses a 1 hour prompt cache window for the main agent, if you leave your computer for over an hour then continue a stale session, it's often a full cache miss. To improve this, we have shipped a few…

Boris, you're seeing a ton of anecdotes here and Claude has done something that has affected a bunch of their most fervent users. Jeff Bezos famously said that if the anecdotes are contradicting the metrics, then the metrics are measuring the wrong things. I suggest you take the anecdotes here seriously and figure out where/why the metrics are wrong.

We are taking it seriously, and are continuing to investigate. We are not trusting the metrics.

Re: Pro Max 5x quota exhausted in 1.5 hours despite moderate usage

#294

Post by ac29 »

coderbants wrote: post #140
SkyPuncher wrote: post #125

I skimmed the issue. No wonder Anthropic closes these tickets out without much action. That’s just a wall of AI garbage. Here’s what I’ve done to mostly fix my usage issues: * Turn on max thinking on every session. It save tokens overall because I’m not correcting it of having it waste energy on dead paths. * keep active sessions active. It seems like caches are expiring after ~5 minutes (especially during peak usage…

Can’t you turn the 1M off with a /model opus (or /model sonnet)? At least up until recently the 1M model was separated into /model opus[1M]

1M context window is still a separate, non-default model in Claude Code and not included with subscriptions (billed at API rates only)

Re: Pro Max 5x quota exhausted in 1.5 hours despite moderate usage

#295

Post by bcherny »

echelon wrote: post #286
rawicki wrote: post #247

Earlier quoted context omitted.

I started seeing "not a malware, continuing" in almost every reply since around 2 weeks ago. Maybe you just reintroduced it with some regression? Opus 4.6

1. I've never seen this. Is there a config option to unhide it if it's happening? Is this in Claude Code? Does it have to be set to verbose or something? 2. Can we pay more/do more rigorous KYC to disable it if it's active?

This warning is not enabled for modern models. No action needed. I'm digging into the report above as soon as they're able to /feedback.

Re: Pro Max 5x quota exhausted in 1.5 hours despite moderate usage

#296

Post by weavie »

throwaway2027 wrote: post #177
weavie wrote: post #90

How good are local LLMs at coding these days? Does anyone have any recommendations for how to get this setup? What would the minimum spend be for usable hardware? I am getting bored of having to plan my weekends around quota limit reset times...

Some claim that some of the recent smaller local models are as good as Sonnet 4.5 of last year and the bigger high-end models can be as almost as good as Claude, Gemini and Codex today, but some say they're benchmaxed and not representative. To try things out you can use llama.cpp with Vulkan or even CPU and a small model like Gemma 4 26B-A4B or Gemma 4 31B or Qwen 3.5 35-A3B or Qwen3.5 27B. Some of the smaller quant…

Thank you, I'll look into it. For someone who is used to just working with second hand thinkpads, this stuff gets expensive fast!

Re: Pro Max 5x quota exhausted in 1.5 hours despite moderate usage

#297

Post by samuelknight »

bcherny wrote: post #210

Hey all, Boris from the Claude Code team here. We've been investigating these reports, and a few of the top issues we've found are: 1. Prompt cache misses when using 1M token context window are expensive. Since Claude Code uses a 1 hour prompt cache window for the main agent, if you leave your computer for over an hour then continue a stale session, it's often a full cache miss. To improve this, we have shipped a few…

Have you considered poking the cache?

When a user walks away during the business day but CC is sitting open, you can refresh that cache up to 10x before it costs the same as a full miss. Realistically it would be <8x in a working day.

Re: Pro Max 5x quota exhausted in 1.5 hours despite moderate usage

#298

Post by pxc »

holoduke wrote: post #244
pxc wrote: post #64

Earlier quoted context omitted.

Translation generally works well with very small models compared to the frontier LLMs. You can definitely run a model on your own hardware for this.

Maybe words. But quality texts in even with opus not perfect. But good enough.

For short texts, the translation I usually want the most is fast translation, and local models are actually great for this.

But for high-ish quality translations of substantive texts, you typically want a harness that's pretty different from Claude Code. You want a glossary of technical terms or special names, a structured summary of the wider context, a concise style guide, and you have to chop the text into chunks to ensure nothing is missed. Even with super long context models, if you ask them to translate much at once they just translate an initial portion of it and crap out.

Are you using it for localization or short strings of text in an app? I wonder what you can do to get better results out of smaller models. I'm confident there's a way.

Re: Pro Max 5x quota exhausted in 1.5 hours despite moderate usage

#299

Post by dkersten »

bcherny wrote: post #210

Hey all, Boris from the Claude Code team here. We've been investigating these reports, and a few of the top issues we've found are: 1. Prompt cache misses when using 1M token context window are expensive. Since Claude Code uses a 1 hour prompt cache window for the main agent, if you leave your computer for over an hour then continue a stale session, it's often a full cache miss. To improve this, we have shipped a few…

Eh you say that every time and yet it keeps happening.

Re: Pro Max 5x quota exhausted in 1.5 hours despite moderate usage

#300

Post by tvbusy »

omosubi wrote: post #160
palata wrote: post #150

Earlier quoted context omitted.

> We may very well look back on the last couple years as the golden era of subsidized GenAI compute. Looks like enshittification on steroids, honestly.

Getting $5000 worth of product essentially free and then being told to pay is not enshittification.

We have seen this before. Companies using VC money to take over the market and then increase prices. In the end, we're worse off without these scumbags but some will still sing that we got free service do it's bot enshitification.
Post reply on HN