Perplexity ships Decider decision models that return typed probabilities instead of text
antoine_chaffin · x · 2026-11-13
Perplexity team member Antoine Chaffin announced that Decider V1.3 and its smaller, cheaper sibling Decider Flash V1.3 are now available on OpenRouter (free for one week), Hugging Face (Apache license), and Perplexity's API.
These "decision models" don't generate text. Instead, they read content passed as state (text, JSON, or images) and return typed, probabilistic answers in a single request: probability of yes for yes/no questions, probabilities per option for choice questions (up to 255 options), or per-level probabilities plus an expected score for ordered rubrics. They target high-volume classification, routing, and moderation, where application code thresholds the returned numbers rather than parsing chat replies.
The release also adds a locate feature on the Extractions route, pulling text and list fields with source spans and confidences. The team says more features are still in the works.
More from Models
- Chinese major models reportedly trained on Blackwell-class chips, though scale disappoints — teortaxesTex · 2026-11-13
- One user burned $4,988 of API-priced tokens on a $200 Claude subscription last month — jonathan_wilke · 2026-11-13
- Gemini 4 Hydrogen tops Google models on AA-Robotics v0.1, ranking fourth overall — ArtificialAnlys · 2026-11-13
- 4,689-trial agent test: DeepSeek V4 argues its way into failures, Gemini 3.8 Flash fails silently — Kindly-Bag763 · 2026-11-13
- Analyst: DeepSeek V4's Flash variant is winning, and a hyper-efficient 2B V5 could be the real play — teortaxesTex · 2026-11-13
- Interaction-as-a-Judge beats code rewards: +6.7pp win rate, 35.7% less code bloat — arena · 2026-11-13