undefined

points

[-]

That explains things. Im getting this: API Error: 400 {"error":{"message":"Budget has been exceeded! Current cost: 271.29866200000015, Max budget: 200.0","type":"budget_exceeded","param":null,"code":"400"}}

So I completetly ran out of tokens and haven’t even used it at all for the past couple of days, and last week my usage was very light. Let me scratch that, all my usage has been very light since I got this plan at work. It’s a an enterprise subscription I believe, hard to tell since it doesn’t connect directly to Anthropic, rather it goes through a proxy on Azure.

Im not liking this at all and all, so flaky and opaque. Not possible to get a breakdown on what the usage went on, right? Do we have to contact Anthropic for a refund or will they restore the bogus usage?

by nixpulvis2 hours ago|

prev|

[-]

This is a serious problem with the fact that it's nearly impossible to understand what a "token" is and how to tame their use in a principled way.

It's like if cars didn't advertise MPG, but instead something that could change randomly.

by clawfund1 hours ago|

parent|

[-]

The opacity isn't accidental. When customers can't forecast cost, they can't build an internal business case to switch, which is more durable than lock-in from product quality. The first AI coding vendor to offer deterministic $/task pricing will own enterprise procurement conversations. Right now the entire market is hiding behind 'tokens' specifically to avoid that comparison.

by claw-el1 hours ago|

parent|

prev|

[-]

Also, certain models are more verbose than the others. We are basically at the mercy of a model who likes to ramble a lot.

by konfusinomicon37 minutes ago|

parent|

[-]

im fiarly certain the knob on the machine that controls length of redundant comments and docblocks is cranked to 11. it makes me curious how much of their bottom line is driven by redundant comment output.

by amitprasad1 hours ago|

parent|

prev|

[-]

Relevant post: https://modal.com/blog/dollars-per-token-considered-harmful

(disclaimer: I work with the author)

by nixpulvis1 hours ago|

parent|

[-]

I completely agree that requests are what should be charged for. But I think there are two things, given that requests aren't all going to cost the same amount:

1. Estimate free invoicing the requests and letting users figure it out after the fact. 2. Somehow estimating cost and telling users how much a request will cost.

We have 1, we want 2.

by uoaei2 hours ago|

parent|

prev|

[-]

Like if cars measured fuel efficiency or range using the knobs in the tread on your tire.