Google launches Gemini Decisions API: $0.10/1M input tokens, zero output cost
rseroter · x · 2026-11-12
Google introduced the Gemini Decisions API, powered by a new Gemini Decision 1 model built for fast routing, tool selection, safety checks and grading inside apps and agents—without the latency or output-token cost of full text generation.
Key specs: up to 1,000 questions evaluated per request, $0.10/1M input tokens with $0 output tokens, and one shared input spanning text, images, audio, video and PDFs.
Google also applied it to Stitch: a new rapid design experiment turns a rough prompt into a clickable multi-screen flow in about 5 seconds, targeting early-stage design ideation.
Related event: Google Launches Gemini Decisions API with Gemini Decision 1 Model(24 posts)→
More from Models
- VulnCheck audit: only 0.57% of Anthropic Glasswing's 37,966 findings ever fixed — WeldPond · 2026-11-12
- On PostTrainBench, AI post-training now beats the human-tuned SmolLM3-3B from the same base — maksym_andr · 2026-11-12
- Claude flags mispronunciations in video audio it claims it can't hear — RileyRalmuto · 2026-11-12
- Heavy Users Say Claude's Safety Rails Block Even Basic Chemistry Work, Forcing Fallback to Opus — rbhar90 · 2026-11-12
- Community teams release wave of abliterated Qwen models for 8–24GB local GPUs — udmrzn · 2026-11-12
- Redditor claims Qwen 4 Max slots between DeepSeek v4.1 and Kimi 3.1 Pro in output quality — power97992 · 2026-11-12