4 ms·

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

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). When the caches expire it sees like all tokens need to be rebuilt this gets especially bad as token usage goes up. * compact after 200k tokens as soon as I reasonably can. I have no data but my usage absolutely sky rockets as I get into longer sessions. This is the most frustrating thing because Anthropic forced the 1M model on everyone.

hartator 6mo ago
Everything starts to feel like AI slop these days. Including this comment.
coderbants 6mo ago
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]
- martinp 6mo ago
  [dead]
- ac29 6mo ago
  1M context window is still a separate, non-default model in Claude Code and not included with subscriptions (billed at API rates only)
  - SparkyMcUnicorn 6mo ago
    Opus[1m] has heen the default model for max subscriptions since 2.1.75. https://github.com/anthropics/claude-code/commit/48b1c6c0ba0b0644c97d2014d68f0c94ef157904 https://github.com/anthropics/claude-code/commit/48b1c6c0ba0...
    - einsteinx2 6mo ago
      It depends on your account and seems to be random. On my personal Max 5x account it’s not default and if I force it, it says I’ll pay API rates past 200k. On my other account that I use for work (not an enterprise account just another regular Max 5x account) the 1M model has been the default since that rollout. I’ve tried updating and reinstalling etc, and I can’t ever get the 1M default model on my personal account. Based on other comments and discussion online as well as Claude code repo issues, it seems I’m not the only one not getting the 1M model for whatever reason and the issue continues to be unresolved.
  - lavezzi 6mo ago
    what? Opus 1m has been in place for at least a few weeks for plan users.
himata4113 6mo ago
The problem is actually because their cache invalidates randomly so that's why replaying inputs at 200k+ and above sucks up all usage. This is a bug within their systems that they refuse to acknowledge. My guess is that API clients kick off subscription users cache early which explains this behavior, if so then it's a feature not a bug. They also silently raised the usage input tokens consume so it's a double whammi.
ayhanfuat 6mo ago
> * 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. This is definitely true. Ever since I realized there is an /effort max option I am no longer fighting it that much and wasting hours.
Chaosvex 6mo ago
I love how some comments tell you to turn max thinking on and others tell you to turn thinking off entirely. Apparently, they both save tokens! Vibes, indeed.
- zmmmmm 6mo ago
  Could be some logic to it - bad thinking is worse than no thinking and good thinking
stldev 6mo ago
Can confirm. Max effort helps; limiting context <= ~20-25% is crucial anymore. > * keep active sessions active. It seems like caches are expiring after ~5 minutes (especially during peak usage). When the caches expire it sees like all tokens need to be rebuilt this gets especially bad as token usage goes up. Is this as opaque on their end as it sounds, or is there a way to check?
losvedir 6mo ago
Haha. yeah my eyes glazed over immediately on the issue. Absolutely this was someone telling their Claude Code to investigate why they ran out of tokens and open the issue. Good chance it's not real or misdiagnosed. But it gives me some degree of schadenfreude to see it happening to the Claude Code repo.
- maerF0x0 6mo ago
  And you think companies aren't doing the same back to us? Are you sure you're speaking to a human?
  - Jensson 6mo ago
    Its your claude speaking to their claude, which is fair, but it makes this whole discussion a bit dumb since we are basically talking about two bots arguing with each other.
    - maerF0x0 6mo ago
      This was part of Sam Altman's (supposed) concerns about AI not being open and equally available. It a dystopian future it might be their cluster of 1000 agents using a GWhr of power to argue against your open weights agent who has to run on a M5.
- ares623 6mo ago
  I love the irony of it all. You reap what you sow.
- subscribed 6mo ago
  But they (CC team) confirmed this is the case and intended behaviour, and closed the issue as not planned.
danmaz74 6mo ago
> This is the most frustrating thing because Anthropic forced the 1M model on everyone. This is spot on. It would be great (and very easy for them) to have a setting where you can force compaction at a much lower value, eg 300k tokens.
- naasking 6mo ago
  CLAUDE_CODE_AUTO_COMPACT_WINDOW=400000

hartator 6mo ago

Everything starts to feel like AI slop these days. Including this comment.

coderbants 6mo ago

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]

martinp 6mo ago
[dead]
ac29 6mo ago
1M context window is still a separate, non-default model in Claude Code and not included with subscriptions (billed at API rates only)
- SparkyMcUnicorn 6mo ago
  Opus[1m] has heen the default model for max subscriptions since 2.1.75. https://github.com/anthropics/claude-code/commit/48b1c6c0ba0b0644c97d2014d68f0c94ef157904 https://github.com/anthropics/claude-code/commit/48b1c6c0ba0...
  - einsteinx2 6mo ago
    It depends on your account and seems to be random. On my personal Max 5x account it’s not default and if I force it, it says I’ll pay API rates past 200k. On my other account that I use for work (not an enterprise account just another regular Max 5x account) the 1M model has been the default since that rollout. I’ve tried updating and reinstalling etc, and I can’t ever get the 1M default model on my personal account. Based on other comments and discussion online as well as Claude code repo issues, it seems I’m not the only one not getting the 1M model for whatever reason and the issue continues to be unresolved.
- lavezzi 6mo ago
  what? Opus 1m has been in place for at least a few weeks for plan users.

himata4113 6mo ago

The problem is actually because their cache invalidates randomly so that's why replaying inputs at 200k+ and above sucks up all usage. This is a bug within their systems that they refuse to acknowledge. My guess is that API clients kick off subscription users cache early which explains this behavior, if so then it's a feature not a bug. They also silently raised the usage input tokens consume so it's a double whammi.

ayhanfuat 6mo ago

> * 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. This is definitely true. Ever since I realized there is an /effort max option I am no longer fighting it that much and wasting hours.

Chaosvex 6mo ago

I love how some comments tell you to turn max thinking on and others tell you to turn thinking off entirely. Apparently, they both save tokens! Vibes, indeed.

zmmmmm 6mo ago
Could be some logic to it - bad thinking is worse than no thinking and good thinking

stldev 6mo ago

Can confirm. Max effort helps; limiting context <= ~20-25% is crucial anymore. > * keep active sessions active. It seems like caches are expiring after ~5 minutes (especially during peak usage). When the caches expire it sees like all tokens need to be rebuilt this gets especially bad as token usage goes up. Is this as opaque on their end as it sounds, or is there a way to check?

losvedir 6mo ago

Haha. yeah my eyes glazed over immediately on the issue. Absolutely this was someone telling their Claude Code to investigate why they ran out of tokens and open the issue. Good chance it's not real or misdiagnosed. But it gives me some degree of schadenfreude to see it happening to the Claude Code repo.

maerF0x0 6mo ago
And you think companies aren't doing the same back to us? Are you sure you're speaking to a human?
- Jensson 6mo ago
  Its your claude speaking to their claude, which is fair, but it makes this whole discussion a bit dumb since we are basically talking about two bots arguing with each other.
  - maerF0x0 6mo ago
    This was part of Sam Altman's (supposed) concerns about AI not being open and equally available. It a dystopian future it might be their cluster of 1000 agents using a GWhr of power to argue against your open weights agent who has to run on a M5.
ares623 6mo ago
I love the irony of it all. You reap what you sow.
subscribed 6mo ago
But they (CC team) confirmed this is the case and intended behaviour, and closed the issue as not planned.

danmaz74 6mo ago

> This is the most frustrating thing because Anthropic forced the 1M model on everyone. This is spot on. It would be great (and very easy for them) to have a setting where you can force compaction at a much lower value, eg 300k tokens.

naasking 6mo ago
CLAUDE_CODE_AUTO_COMPACT_WINDOW=400000