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...
| Field | Per token (USD) | Per million |
|---|---|---|
| prompt | 0.00000015 | $0.150 |
| completion | 0.0000006 | $0.600 |
| input_cache_read | 0.000000075 | $0.075 |
{
"temperature": null,
"top_p": null,
"frequency_penalty": null
}