Perplexity Decider v1.3 Adds Locate Extraction

// 28m agoOPENSOURCE RELEASE

Perplexity Decider v1.3 Adds Locate Extraction

Perplexity open-sourced its latest decision models under Apache licensing, adding Locate for extracting exact text spans and character positions. Both models are free on OpenRouter for one week and support classification, routing, moderation, reranking, and structured evaluation.

// ANALYSIS

Perplexity is pushing decision models beyond simple classification toward fast, composable infrastructure for AI applications.

–Locate enables named-entity recognition, PII detection, and extractive question answering without generative parsing
–Decider v1.3 returns calibrated probabilities for binary, multiclass, and ordered-score decisions
–Flash offers a smaller, distilled option for high-volume, latency-sensitive workloads
–Open weights let developers self-host, while OpenRouter lowers the barrier to testing the models immediately
–The combination of locate and decision outputs could simplify routing and quality-control loops inside agents

// TAGS

decider-v1.3decider-flash-v1.3llmopen-weightsopen-sourcestructured-outputsearchreranker

DISCOVERED

28m ago

2026-11-12

PUBLISHED

51m ago

2026-11-12

RELEVANCE

9/ 10

AUTHOR

AravSrinivas

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

UPDATE44m ago

Framer Agent Generates Images, Shaders

Framer’s AI Agent can now turn natural-language prompts into generated imagery and interactive shader effects, including pointer-driven smudges, halftones, CRT visuals, bokeh, and pencil styles. The workflow aims to produce polished web interactions without manual code or After Effects.

LAUNCH56m ago

Daybreak Red Team automates continuous pen tests

OpenAI’s Daybreak Red Team puts coordinated AI agents to work testing web applications, validating vulnerabilities, and checking whether patches actually stop exploits. The early-access service targets organizations that need continuous offensive security coverage without maintaining a dedicated red team.

BENCHMARK1h ago

KhalilBench Turns Replies Into Scores

KhalilBench is a tongue-in-cheek benchmark that scores @kamkmgamer’s replies to model tests. It reframes model evaluation around personality, humor, and social impact rather than standardized task accuracy.