Chrome Decisions API: the prompts, limits, engines and code reviews behind DecisionModel
DEJAN
Chrome’s DecisionModel API code landed in Chromium between 6 and 8 October 2026 as an on-device web API for evaluating closed questions about a text and returning per-answer probabilities. Chrome’s EmbeddingGemma engine runs on an 8-question batch in about 9.6 seconds on CPU. Review findings and experiments show option probabilities are computed via softmax over cosine similarities with limits on questions, options, and rating levels, after which Chrome now scores each option against its own session to avoid earlier-answer leakage and blocks model downloads without a user gesture from 8 October.
Why it matters
An investigation of Chrome's experimental Decisions API compares the explainer with implementation code and review comments. It surfaces details like slow Gemma 4 timings, position bias, context limits, probability semantics, and differences from MediaPipe's DecisionMaker.
Related stories
AWS Strands Labs Releases Strands Decider 2B: An Open Source Decision Model That Picks Options in About 115 ms
MarkTechPost · 1 week ago ·
48
OpenAI Decisions API: faster typed answers for apps
OpenAI Developers · 2 days ago ·
20
AWS launches a local answer to TypeSafe’s Jev decision model
The New Stack · 1 week ago ·
36