RightNow System Zero is a streaming decision model API: you send your inputs with your questions, and it returns decisions with per-option probabilities while the input is still arriving. It launched on October 30, 2026, and this post is based on the official website and pricing docs, checked on November 1, 2026.
What System Zero does
One model takes eight input kinds: text, JSON, robot state, codebases, PDF, images, voice, and video (preview), across five file formats including PNG, JPEG, WAV, PDF, and MP4. The decision updates as each part arrives, so there is no need to wait for the whole input. Up to 1,024 questions can ride in a single call, with 38,000 tokens of standard context and an optional 1M context window set in the console.
The official site publishes decision benchmarks with verified provenance, placing System Zero among dedicated decision models, and notes its own measurements are not an Artificial Analysis evaluation. Sessions get routing affinity, and the decision stream carries per-option probabilities rather than a single bare answer.
Who it is for
System Zero suits teams building decision layers into products: support and operations automation that must pick an option with a confidence attached, robotics and agent systems feeding state and code, and any pipeline where waiting for the full input costs too much time.
Pricing
Pricing is usage based, per the official pricing page updated October 29, 2026. Uncached text input is $0.19 per million tokens and cached text input is $0.019; an answered question costs $0.00013 and an answered call $0.000055; images are $0.011 each, audio $0.00036666667 per started second with a 50-second minimum per file, PDFs $0.011 per page with a $0.0055 minimum per file, and output tokens are free. Refused questions are not charged. Credits top up from a $5 minimum, with starting credit subject to eligibility checks. Prices may change, so check the official website for the latest pricing: https://www.rightnowai.co/.
