Anthropic and Google cut API prices by 50% for batch work. Here is what you give up, what you save at real volumes, and which workloads should never use it.
The exact formula behind every LLM API bill, a worked example, and the three mistakes that make estimates worthless.
A 5x surcharge hiding at 100K tokens, prices that double on a calendar date, thinking tokens you never see — the fine print that actually decides your API bill.
Three providers, three pricing philosophies. Where each one is cheapest, where the fine print lives, and why the answer flips with your workload.
Cached input tokens cost a tenth — or a twentieth — of fresh input on the major APIs. For prompt-heavy agents, that is the difference between a viable product and a bonfire.