openai/gpt-oss-120b
deepinfra
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...
- Context window
- 131,072 tokens
- Input cost
- —
- Output cost
- —
- Latency (p50)
- —