Astra on UltraFast is a delight. Is Claude holding back because of compute?
retained text
With ChatGPT throwing resets left, right and center I thought it was a good time to try to run everything on Ultra Fast. I thought Sam Altman was exagerating when he said once you try it you will never go back. It's incredibly nice. The context switching headaches I get are gone because it's so fast I can stay within one chat. The problem is usage of course. I run through my weekly usage in a few hours. I now use the 1.5 speed as default (using Sol medium, but bump to Astra when needed). I'm now paying $500 at ChatGPT for their Pro500 plan. I find it odd I don't have this option in Claude. I would happily pay $500 or a $1000 for more usage and faster speed. I'm not even looking for linear or exponentially more. But $100/mo. just seems limiting. Give me a $500 plan with 3x the usage or $1,000 plan for 5x the usage. Is it simply because Claude don't have any compute left to service everybody? I get that our plans are "subsidized", but I'm not a huge plan of that argument. PS. I can't wait for Anthropic and OpenAI to figure out the speed vs efficiency balance. Going back to 1x speed feels sooooo slow.
document fetched 2026-10-29 00:22 · canonical https://reddit.com/r/ClaudeAI/comments/1xhwl9z/astra_on_ultrafast_is_a_delight_is_claude_holding