Google launches Gemini Decisions API for fast multimodal decisions

It evaluates up to 1,000 questions across text, images, audio, video, and PDFs in a single pass, returning probabilities and confidence scores with 150–250ms end-to-end latency and $0.10 per million input tokens. The model achieves the lowest calibration error across agent routing, reasoning, and multimodal understanding benchmarks. Used internally for dynamic routing, design iteration, emoji selection, and safety checks.

Summarised from the original article by an open-source model, then checked automatically: every number and English name must appear in the source. The original article is authoritative.