> The cost to serve tokens is absolutely profitable

Can you explain why you know better than the analyst at Cursor cited in this article?

iterateoften a day ago | parent | next [-]

Open router is an upper bound of compute cost for the open source models. So people assume that opus and sonnet really isn’t sucking up 10x the resources because open source models aren’t 10x worse. Idk if it’s true or not, but haiku is $5/m tokens and it is much worse than the $2-3/mt models imo

	▲	vbarrielle a day ago \| parent [-]
		Openrouter is a startup, what's the indication it serves token at a profit? It could be serving them at a loss to show growth.

▲

infecto a day ago | parent | prev | next [-]

Can you cite your source of an analyst at Cursor. I read the article and looking through the boatload of links but struggled to find what you are referring to. Ty

▲

noelsusman a day ago | parent | prev [-]

That analyst was talking about subsidizing tokens through the subscription plans, which is a different claim.

	▲	infecto a day ago \| parent [-]
		Ty for sharing and agree. I think there is confusion with some folks in the comments for this post confusing inference profitability and plan profitability. Most plans as we can tell are probably teetering the line of profitability and that’s why we have seen some like Cursor really tighten how many tokens you get.