we need 100x plan

I'm curious why you need that and how did you managed to spend all tokens?

Using 5x plan I'm barely fit between time blocks (usually my tokens ends 15-20 minutes before next time slot).

If I run multiple concurrent agents in my tasks I need 20x then, but never was able to hit 20x limit even with doing a lot of work at the same time on multiple projects...

Hello @morsik !

so my workflow is a bit different from typical usage, I'm building an autonomous agent orchestration system where one command spawns 8-11 parallel agents, each working with full context windows. they read entire codebases, write outputs to disk, then I have verification agents that audit and refine those outputs in loops until they pass quality checks

think of it like this, instead of one conversation back and forth, running a mini engineering team inside cc. one cycle can cascade through 10-20+ agents, each burning through full context. so on the 20x plan, what should be a 5-hour window barely lasts me 45 minutes to an hour (if intensive).

ur 5x and 20x usage makes total sense for single-threaded work or even a few concurrent agents. but when you scale to full orchestration with parallel processing + audit loops + refinement cycles, the token consumption is on a completely different level. that's why i am asking for 100x, some of us are pushing cc beyond single-developer workflows into something closer to an AI operations layer

@Davronov-Alimardon I would see my self using this because as how you explained on how over 10+ agnets would neeed 100x plan

Interesting, I'm using a lot of parallel agents and never hit limits in Max 20. And I've run even 30 agents in parallel on some occasions.

But I guess, Anthropic just wants you to go into API, as subscriptions is too cheap already anyway and this would disrupt infrastructure a lot...

@morsik have u ever batched 30/40 agents in a single request? not 30 separate agents across a session. 30 firing at once from one command, each with full context. and then doing that 5-10+ times in one 5h window? that's where limits hit fast

fair point on API though, you're right. but API pricing adds up quick, i'm burning through $40-50 in the first hour alone at that rate. a fixed 100x tier even at a higher price than i originally suggested would still be more predictable. but yeah, we'll see what Anthropic decides (probably ignore 😄)

@DevFlex-AI appreciate it, seems like there's real demand for this tier beyond just my workflow ))

+1

Especially with how much Opus 4.7 eats tokens

@atlamors Have you faced the issue where Claude code glitches and just loads and then says I hit my limit before I went to the max plan

we definitely need 100x plan

Closing for now — inactive for too long. Please open a new issue if this is still relevant.

And now with dynamic workflows it would be useful... :D

I'm eating my whole weekly limit in 2-3 days of Max20 plan with this new feature... and it actually works great!

Yepp a 100x plan (or 200x if fast mode is included) is needed given how workflows have advanced. 20x is just too restrictive nowadays, gone in less than a day and that's with constant context and model management. Would love to just pay more and be able to use the best model and not care about micro-managing the plan, and rather put all focus into building 🙏 🙏 🙏

I'm thinking about just creating secondary account and buying second subscription... Apparently it's not against ToS, but... it's still kinda grey area :/