Also the biggest models are not always the best depending on what you're doing. 4.6 Sonnet is a decent model that can handle most coding tasks, and even 4.5 Haiku is fine for simple, well-defined tasks.
Using 4.6 Opus for simple things is not only wasting tokens, it's also slower. Sonnet will get a lot of tasks done in half the time for less than half the money
Using 4.6 Opus for simple things is not only wasting tokens, it's also slower. Sonnet will get a lot of tasks done in half the time for less than half the money