OpenAI: gpt-oss-20b (free)

OpenAI's gpt-oss-20b is an open-weight 21B parameter model, freely available under the Apache 2.0 license. This model utilizes a sophisticated Mixture-of-Experts (MoE) architecture, with only 3.6B active parameters per forward pass, making it highly optimized for lower-latency inference and deployable even on consumer or single-GPU hardware. It's trained in OpenAI’s Harmony response format, ensuring high-quality outputs. This powerful model supports a range of advanced capabilities, including reasoning level configuration, fine-tuning, and robust agentic features like function calling, tool use, and structured outputs. With a generous 131K token context window and a 4K token maximum output, gpt-oss-20b is excellent for complex chat interactions and code generation. Best of all, it's completely free to use on Multi AI, offering exceptional value for developers and users alike.