Luckily inference is cheap, and other providers offer efficient models.
It’s only going to get better in the future.
It’s only going to get better in the future.
That’s about 1.45x more expensive than K2.5 from January.
It is around 5x cheaper than GPT 5.5.