發生什麼事
The API handles text, images, audio, video, and PDFs. Google says it is available to try in AI Studio and through the Gemini API, and its Gemini Decisions API: Fast, calibrated multimodal decisions describes the service. The model is described as making fast, calibrated decisions, with up to 1,000 questions in one request (@GoogleAIStudio).
Pricing details came from two different sources. @OfficialLoganK, an individual X account rather than an official account, reported a price of $0.10 per million input tokens and a 256K context window. @OpenRouter said the model was listed on its platform on launch day at the same input price, with output free.
01Sources
Evidence: 🗣 official self-report; not tested by us. 🗣 personal-account self-report; not tested by us.
- @GoogleAIStudio: Google AI Studio official account, posted 2026-11-11 18:15 UTC, announcing Gemini Decisions API.
- Gemini Decisions API: Fast, calibrated multimodal decisions: Google AI Blog, published 2026-11-11 18:00 UTC.
- @OpenRouter: OpenRouter official account, posted 2026-11-11 18:23 UTC, announcing availability and pricing.
- @OfficialLoganK: individual X account, posted 2026-11-11 18:16 UTC, reporting price and context length.