From the source
- Google launched the Gemini Decisions API, a service that answers only three kinds of questions, powered by a new model called Gemini Decision 1.
- Traditional large language models answer a broad range of prompts, while Gemini Decision 1 handles yes-no questions, multiple choice questions and prompts that rate an item on a user-specified scale, which Google says makes it nearly five times cheaper than its entry-level Gemini 3.5 Flash-Lite LLM.
- Prompts pair a question with a dataset of up to 256,000 tokens of text, audio, images, video or a combination, and the model returns terse answers such as "yes" or "no" plus a confidence score that helps applications decide whether to use the output.
Google DeepMind's Min Choi and Felix Yu wrote that the model scores options without token-by-token decoding, keeping median latency at 50-100 ms of model compute and 150-250 ms end-to-end across multimodal inputs and contexts up to 16K tokens.
- Google says Gemini Decision 1 beat Jev and
OpenAI's Decisions API on seven of nine benchmarks, though Jev was slightly more cost-efficient.
- The API costs $0.10 per 1 million input tokens with no charge for output tokens, and a limited free tier is available in Google AI Studio.