What is changing across AI model APIs?
Across 194 public accounts, Flash pricing and multimodal input draw the most praise, and surprise bills and model retirements the most concern. Team accounts raise retirements far more often than individual builders do, and DoorDash calls them a migration cost across hundreds of use cases.
Budget Gemini 3.8 Flash at its standard US$1.50 and US$7.50 from January 1, 2027, twice the introductory price. Batch, Flex, and caching decide how much of that increase reaches your cost per task.
Compare model APIs on cost per completed task. Anthropic and OpenAI now list the same prices at matching tiers, so cache hit rate, batch share, and tokens per task decide the bill.
Also covers Claude API and OpenAI API
