Gemini Decisions API

Why it matters

Google has entered the emerging decision model category alongside Jev and OpenAI with Gemini Decision 1. The new API closely mirrors Jev's design, supporting yes/no questions, multiple-choice questions, and rubric scoring, and it even adopts Jev's term "Noul" for yes/no queries. On pricing, it matches OpenAI's GPT-6 Luna at ten cents per million input tokens, with free output, while Jev stays the cheapest option. The key differentiator is input support: Jev handles only text and JSON, OpenAI handles text and images, and Gemini Decision 1 accepts text, images, audio, video, and PDFs. The author tested it against a personal video, asking whether it contained a dragon or robot and whether it was set in France or the USA. Whereas Jev and OpenAI return absolute binary values for yes/no questions, Google's model returned graded probabilities, showing less certainty in its answers. For developers building decision-style APIs, the story shows Google matching its rivals on features and pricing while pushing multimodal input further than either competitor.