Qwen: Qwen3 4B (free)

Qwen3-4B is a powerful 4 billion parameter dense language model from the Qwen3 series, engineered to excel in both general-purpose and reasoning-intensive tasks. Its innovative dual-mode architecture allows it to dynamically switch between a 'thinking' mode for high-precision logical reasoning and a 'non-thinking' mode for efficient dialogue generation. This makes it exceptionally versatile for a wide range of applications, including multi-turn chat, instruction following, and complex agent workflows. This free-to-use model boasts a generous 40K token context window and a maximum output of 4K tokens, ensuring comprehensive understanding and detailed responses. It supports advanced capabilities like functions, code generation, and streaming, making it a robust choice for developers and users alike. With zero-cost pricing ($0.00 per 1M tokens for both input and output), Qwen3-4B offers an accessible yet powerful AI solution for various computational needs.