featherless-ai/Qwen3.6-35B-A3B-classifier

Hugging Face
VISIONPricing:Input $0.28 / Output $0Concurrent Unit Cost:2Model Size:35BQuant:FP8Context Size:32k Classification model Architecture:Transformer Warm

The featherless-ai/Qwen3.6-35B-A3B-classifier is a 35 billion parameter Qwen-based model designed for structured decision-making and classification tasks. It leverages the Simple Jev classifier API to convert context into structured JSON outputs, making it suitable for applications requiring intent detection, content moderation, relevance scoring, and agent action selection. This model supports both text and image inputs, offering a versatile solution for various classification needs.

Loading preview...

Overview

The featherless-ai/Qwen3.6-35B-A3B-classifier is a 35 billion parameter model built on the Qwen architecture, specifically engineered for classification tasks using the Simple Jev classifier API. This model excels at transforming diverse inputs, including text, conversations, and images, into structured decisions and scores, outputting results directly in JSON format.

Key Capabilities

  • Structured Classification: Converts context into structured decisions, scores, or judgments.
  • Flexible Question Types: Supports choice for selecting candidates, score for rubric evaluation, and noul for yes/no propositions.
  • Multi-modal Input: Capable of processing both text/structured data and image inputs, making it versatile for various applications.
  • Context Handling: Can evaluate multiple questions against shared context, including chat history.
  • Production Ready: Designed for integration into applications via the Featherless API, with a public demo available for testing.

Good For

  • Support Routing: Automatically directing customer inquiries to appropriate departments.
  • Intent Detection: Identifying user intentions in conversational AI or customer service.
  • Content Moderation: Classifying and filtering inappropriate or irrelevant content.
  • Relevance Scoring: Assessing the relevance of information or documents.
  • Agent Action Selection: Determining the next best action for an AI agent based on current state.
  • Visual Classification: Utilizing image context for classification tasks (e.g., identifying objects or scenes).