gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases. It activates 5.1B parameters per forward pass and is optimized...

• ambient
• baseten
• berget
• crusoe
• deepinfra
• groq
• kilo
• nearai
• nebius
• novita
• parasail
• phala
• siliconflow
• together
• wandb