Model Pricing Tracker

What we’re tracking

(updated …)

Our claim is that we are in the subsidiary phase of Artificial Intelligence providers. Providers are scrambling to build as much infrastructure as they can afford to accommodate increased demand down the road as AI becomes more and more ubiquitous. As we saw in ride-sharing and delivery services companies in the early and mid-2010s, prices for these services were artificially low in an effort to gain market share.

These pricing metrics are all publicly-available as a ‘cost per 1m tokens’ on each provider’s pricing pages and documentation. However, this is only half the metric to determine which models are indeed ‘low cost’. Total cost factors in how efficient they are in providing responses, as the price per token multiplied by tokens consumed is the true measure of consumer spend.

Early findings

There is not as much change as I had anticipated. The models become obsolete and retired before they drop in price. As of writing, only OpenAI actually drops prices in advance of a retiring model. Instead, the industry standard tends to be to introduce a new model (at a slightly higher price per 1m tokens) and retire the old one.

Anthropic’s flagship models are consistently the priciest on the board, rivaling Image creation and realtime audio models from Google and OpenAI. On the other side of the spectrum, Amazon’s Titan and Nova models are pennies on the dollar in comparison. It is still early days, and we’re still in an expanding industry. Check back here for details – we scrape everyday and will keep this page up-to-date.