Saksham Dwivedi

AI/ML Developer| Gamer | ambivert | Indian | Interested in - Technology, Entrepreneurship, Crypto and Stock Market

[Saksham Dwivedi @Lackable03]

@thsottiaux my claude finally started using codex as a throwaway agent to delegate the simplest and smallest of tasks to and I ran out of usage in 2hrs , 40% of my weekly gone in 2 hr sessions, while my claude pro can run subagents, give actual worthwhile output

[Saksham Dwivedi @Lackable03]

@heypetar @pcshipp The weekly on both is exactly the same, only the 5hr session is 4x on 200 of the 100

[Saksham Dwivedi @Lackable03]

@ishuagra02 You get the same weekly usage as 5x plan, just the session limit is 4x of the 5x plan

[Saksham Dwivedi @Lackable03]

@zzaba_a @karanC_12 Stop making your grok bot post for you

[Saksham Dwivedi @Lackable03]

@Al41611876 @JustinGorya If ur not running out of 5hr usage, think about switching to 2 5x accounts since weekly limit on 20x and 5x is the same, only difference is 5hr limit is 4x, get a team plan with 2 premium seats...would get you even more usage

[Saksham Dwivedi @Lackable03]

@ishuagra02 Another 100 sub would give you more usage than 200, the weekly on 200 is same as 100 but only 5hr is increased

[Saksham Dwivedi @Lackable03]

@tiborwell @AICITY112 @ParamSiddh Explore subagent, parallel threads and workflows
I did, can work on multiple things at once now and still my usage could only get to 90% for 5hr
Switched from pro to max 3 days ago

[Saksham Dwivedi @Lackable03]

@laopai1990 Yeah but the output costs more, and thats for under 100k context which coding agents can easily surpass

[Saksham Dwivedi @Lackable03]

@JamesMalsawm High/medium if im on pro while looking over all the edits and everything
Extra with subagents if im on max then review the pr and test

[Saksham Dwivedi @Lackable03]

@Heyy_Meet Got claude max today from work :)

[Saksham Dwivedi @Lackable03]

@gang_kerbal @Its_lakshya_ai Kinda like how Rolls Royce only markets in air-travel conventions

[Saksham Dwivedi @Lackable03]

@bridgebench When are you gonna update usagebench

[Tetris @darkshadow34341]

@clmt @antigravity Fuck Gemini. I’ll stick to GPT Sol and Opus. Your model is trash anyways

[Saksham Dwivedi @Lackable03]

What the 1M output length is being used for lol
Gemini at it again

[Saksham Dwivedi @Lackable03]

@essjaykay755 @theo How did you get these logs?

[Saksham Dwivedi @Lackable03]

@ClaudesRH Less goo

[Saksham Dwivedi @Lackable03]

@reasonmachines When coming on windows?
And no benchmarks for 6.1 sol?

[Saksham Dwivedi @Lackable03]

@Rationalop9 @grok @Winter_arc26 You can go to their unify endpoint and unify both the subscriptions into one (if ur paying more you'll get a refund )

[Saksham Dwivedi @Lackable03]

@theo What are the chances of my claude being banned if I use t3

[Saksham Dwivedi @Lackable03]

@bridgebench When are you going to retest usage bench