Perplexity launches Decisions API: probability-based answers at $0.04/M input tokens
denisyarats · x · 2026-10-02
Perplexity has launched the Decisions API, powered by its decision model pplx-decider-v1-27b. Instead of generating text, it reads text, JSON, or images like a multimodal LLM and returns typed answers with probabilities that software can consume directly.
- Three question types: noul (yes/no, returns probability of yes), choice (picks among custom options, returns per-option probabilities plus the most likely), and score (rates content on an ordered rubric, returning per-level probabilities plus an expected score)
- Designed for classification, routing, and rubric grading where you'd otherwise prompt a chat model and parse its reply — numbers can be thresholded directly, with no output parsing
- Pricing: $0.04 per million input tokens, output tokens free
- The model doesn't write replies, generate code, or explain reasoning; your code does the reasoning with the returned numbers
Related event: Perplexity Open-Sources pplx-decider-v1-27b Decision Model(8 posts)→
More from coding & agent
- Berliner study: existing user-simulator metrics catch only 5-6 of 10 injected failures — geoffwolfe · 2026-11-22
- MCP lead maintainer solicits feedback while also working on Claude Code — dsp_ · 2026-11-22
- DeepORG open-sources an AI agent that runs across your org's databases, APIs and code — dosco · 2026-11-22
- If an agent can still move money after its prompt says no, the boundary is in the wrong place — ExecLayer_io · 2026-11-22
- AutoRAGBench: LLM agents build RAG systems beating the best published human designs — omarsar0 · 2026-11-22
- Stop @-ing your agent in Slack — let Jev decide when it should respond — DavidWells · 2026-11-22