Perplexity · Decision Models
Perplexity Decider v1.3 (pplx-decider-v1.3 and pplx-decider-flash-v1.3)
Perplexity has released Decider v1.3 and Decider Flash v1.3, open-weight decision models that return probabilities for yes/no, multiple-choice and score questions about text or image input rather than generated text. The new Locate feature extracts exact text spans by label or instruction, aimed at NER, PII detection and extractive QA. Perplexity reports 91.6% mean accuracy on 63 query-classification tasks and 83.99 mean nDCG@10 on retrieval, but all benchmark figures are self-reported. Weights are Apache 2.0 on Hugging Face, and the models are also available through the Perplexity Decisions API and OpenRouter.
Open the source ↗Similar Language Models releases
- EmbeddingGemma 2
Google DeepMind · Multimodal Embeddings · Week 42, 2026
- Gemma Translator
Google Creative Lab · Offline Voice Translator · Week 41, 2026
- Ternary Bonsai 2 27B
PrismML · Ternary Quantisation · Week 39, 2026
- Tiny Aya Global
Cohere Labs · Multilingual · Week 8, 2026
- Ming-Lite-Omni
inclusionAI · Multilingual · Week 36, 2025
- LLMON: LLM-Native Markup Language (Granite 4.1 LLMON previews)
IBM Granite (ibm-granite) · Structured Prompt Format · Week 47, 2026
Published by Quanta AI Labs, an AI agency in India. Browse every week in AI News Weekly.