cemig-temp/qwen3-convo1.1-if

Qwen3-4B: A Versatile Language Model with Dynamic Reasoning

Qwen3-4B is a 4 billion parameter causal language model from the Qwen series, designed for advanced reasoning, instruction-following, and agentic tasks. It introduces a novel capability to dynamically switch between two operational modes:

Key Capabilities

• Thinking Mode: Engages for complex logical reasoning, mathematics, and code generation, significantly enhancing performance in these areas compared to previous models.
• Non-Thinking Mode: Optimized for efficient, general-purpose dialogue and conversational tasks, aligning with the functionality of Qwen2.5-Instruct models.
• Superior Human Preference Alignment: Excels in creative writing, role-playing, and multi-turn dialogues, providing a more natural and engaging user experience.
• Advanced Agent Capabilities: Demonstrates leading performance among open-source models in integrating with external tools for complex agent-based tasks.
• Multilingual Support: Supports over 100 languages and dialects with strong multilingual instruction following and translation abilities.
• Extended Context Length: Natively handles up to 32,768 tokens, with validated performance up to 131,072 tokens using the YaRN method for long text processing.

When to Use This Model

• Complex Problem Solving: Ideal for tasks requiring deep logical reasoning, such as mathematical problems or code generation, by leveraging its 'thinking mode'.
• Interactive Applications: Suitable for chatbots, creative writing, and role-playing scenarios due to its strong human preference alignment and multi-turn dialogue capabilities.
• Agent-Based Systems: Excellent for applications requiring tool integration and autonomous task execution.
• Multilingual Applications: Highly effective for global applications needing robust multilingual instruction following and translation.