@NeverSinkDev Opus 5.5 is so damn good, you barely can hit the limits and the compute and token efficency will be getting better and better over time, so its a matter of time, that they need to adjust to make any money. If you can pay $20 bucks and get near infinite frontier models 🫣
@theo opus 5.5 on xhigh cant even deplete my 5h limit with 3-4 tasks running simultanously, I lost like 30% weekly since release and using like crazy. Astra depleted one weekly limit in 2 days.
@pippitofficial worked 1day now its back to normal credit usage, is this only bait? Got support and a „senior agent“ will respond in 1-5 days, what a joke.
@theo my weekly limit dropped like 2% with 2-4 parallel tasks on luna/sol for about 4-5 hours non stop, holy is this a game changer. the same btw. for opus5.5. (20x plan).
@Im_IrushiK I see the same thing with api access skills, it uses for example https://t.co/vmaBxbcVkJ 100x times then it switches to another one and says it isn‘t setup and it can‘t render. Then you tell to use https://t.co/vmaBxbcVkJ and it tries to use the api without apikey 💀
@Kappaemme1926 running Xcode(IDE) with a local llm like qwen etc. already consumes most of the 48 gigs. So 64 Gb would be my sweetspot to run local llm with claude/codex and any IDE/editor/tool to have enough backup.