I'm on subscription but I had Claude add it up: 504M input tokens (90% cached), 7.6M output tokens. 90+% of that was Sonnet calls to label the comments, all of the actual work, the app and deploy etc was ~1M tokens of Opus 5.5, I compacted context once. i.e. very small project by modern standards.
I wish someone would make a decision model that’s as good as sonnet but like 10-100x cheaper. Because for this project which is a pretty small dataset you’re spending ~$300 on sonnet usage. like jev can probably do this task but it’s not smart enough to reliably be used
GLM 5.3 Flash is incredibly cheap and imho as good as or better than Sonnet. I used it recently to fix an old server that my family had running serving some websites, and by fix it I mean I got it to log in over ssh, figure out what's running there, copy it out, log in to a brand new droplet on DO, set up a fresh instance of apache with everything patched up to latest, copy that ancient content over, make sure it runs with latest PHP which required several rounds of fixes, optimize the system so it runs on the smallest 512MB droplet, harden the system so that ideally no one ever needs to look at it for another 10 years - it took about an hour, and cost a grand total of 4 american.... _cents_
I use Opus at work since we have a nearly unlimited AI budget, but for personal projects I just use GLM 5.3 and I don't see much difference between that and Opus.
I think most people could consider the cost as being whatever percentage of the subscription quota was used multiplied by the cost of the subscription.
I'm on subscription but I had Claude add it up: 504M input tokens (90% cached), 7.6M output tokens. 90+% of that was Sonnet calls to label the comments, all of the actual work, the app and deploy etc was ~1M tokens of Opus 5.5, I compacted context once. i.e. very small project by modern standards.
I wish someone would make a decision model that’s as good as sonnet but like 10-100x cheaper. Because for this project which is a pretty small dataset you’re spending ~$300 on sonnet usage. like jev can probably do this task but it’s not smart enough to reliably be used
GLM 5.3 Flash is incredibly cheap and imho as good as or better than Sonnet. I used it recently to fix an old server that my family had running serving some websites, and by fix it I mean I got it to log in over ssh, figure out what's running there, copy it out, log in to a brand new droplet on DO, set up a fresh instance of apache with everything patched up to latest, copy that ancient content over, make sure it runs with latest PHP which required several rounds of fixes, optimize the system so it runs on the smallest 512MB droplet, harden the system so that ideally no one ever needs to look at it for another 10 years - it took about an hour, and cost a grand total of 4 american.... _cents_
I use Opus at work since we have a nearly unlimited AI budget, but for personal projects I just use GLM 5.3 and I don't see much difference between that and Opus.
GLM is really good, but sonnet outclasses it. Anthropic really cooked with sonnet 5+
Yeah i'm pretty sure this is doable multiple times on a 20 dollar subscription, just wow
So it's unknowable how much it cost?
It'd cost ~$250 via API (ie. paying current API pricing per token).
Source: copy pasted the tokens into a chatbot and told it to figure it out.
But hardly anybody pays API pricing and this was made with a sub..
So the cost was zero.
Uh, with some fancy accounting..
I think most people could consider the cost as being whatever percentage of the subscription quota was used multiplied by the cost of the subscription.