Models / QwenQwen / / Qwen3-Next-80B-A3B-Thinking API
Qwen3-Next-80B-A3B-Thinking API

This model is not currently supported on Together AI.
Visit our Models page to view all the latest models.
Advanced Reasoning Engine:
Qwen3-Next Thinking features the same highly sparse MoE architecture but specialized for complex reasoning tasks. Supports only thinking mode with automatic tag inclusion, delivering exceptional analytical performance while maintaining extreme efficiency with 10x+ higher throughput on long contexts and may generate longer thinking content than predecessors.
Qwen3-Next-80B-A3B-Thinking API Usage
Endpoint
How to use Qwen3-Next-80B-A3B-Thinking
Model details
Architecture Overview:
• 48 layers with 2048 hidden dimension and hybrid layout pattern
• 512 total experts with 10 activated and 1 shared expert per MoE layer
• Multi-token prediction mechanism for faster analytical processing
Thinking Mode Optimization:
• Supports only thinking mode with automatic tag inclusion
• Specialized post-training for complex reasoning chains on 15T tokens
• May generate longer thinking content for comprehensive analysis
Performance Characteristics:
• 262K native context length, extensible to 1M tokens with YaRN scaling
• More than 10x higher throughput on contexts over 32K tokens
• SGLang and vLLM deployment support with Multi-Token Prediction
Prompting Qwen3-Next-80B-A3B-Thinking
Applications & Use Cases
Research & Analysis:
• Scientific research and hypothesis generation with detailed reasoning
• Complex data analysis and pattern recognition
• Academic writing and literature review with analytical depth
Problem Solving:
• Engineering design challenges requiring multi-step analysis
• Strategic business planning and decision-making support
• Mathematical problem solving and proof generation
Professional Applications:
• Legal case analysis and argument construction
• Medical diagnosis assistance with reasoning transparency
• Financial modeling and risk assessment with detailed rationale