AI News

Perplexity Open-Sources Decider v1.3 and Decider Flash v1.3 Decision Models

Aravind SrinivasUpdated Nov 13 07:45

AI summary

Perplexity has open-sourced two decision models, Decider v1.3 and Decider Flash v1.3, and is offering them free on OpenRouter for one week. v1.3 adds a locate capability that returns matching text and character offsets based on a description, suitable for batch decision-making as well as evidence and entity extraction. The company says it significantly outperforms the best general-purpose extraction models on 17 out-of-domain NER, PII, and extractive QA tasks, and achieves 6.54 percentage points higher accuracy than the currently used classifier on its own 63 production classification tasks. The weights have been released on Hugging Face.

Why read this

AI score 58

The original article provides the open-source weights of the new version, a free trial entry point, and comparative data on two types of tasks, which can be used to judge whether extraction-type tasks are worth migrating.

Topics

Tags

Quoted post

Perplexity Developers

@perplexitydevs

We're rolling out pplx-decider-v1.3 and pplx-decider-flash-v1.3, the latest version of our open-weight decision models. v1.3 adds locate, a new feature that extracts spans based on type or instructions. Now available in our Decisions API, @huggingface, and @OpenRouter. https://t.co/bsTSl2tHSo

View original

We're rolling out pplx-decider-v1.3 and pplx-decider-flash-v1.3, the latest version of our open-weight decision models. v1.3 adds locate, a new feature that extracts spans based on type or instructions. Now available in our Decisions API, @huggingface, and @OpenRouter. https://t.co/bsTSl2tHSo

View original post

Body

We're open-sourcing our latest versions of Decision Models: Decider v1.3 and Decider Flash v1.3, and making them free on @OpenRouter for a week.

The new capability in v1.3 models is "locate". This lets you describe what to find, and the model returns the matching text with character offsets.

This is useful for batch-mode decision + evidence/entity extraction. On 17 out-of-domain NER, PII and extractive QA tasks, Decider models significantly outperform the best general extraction model.

And, on 63 of our own production classification tasks, v1.3 is 6.54 points more accurate than the classifiers we run today.

Weights are open on Hugging Face: https://t.co/6olEOrOq4y

Source

Aravind Srinivas