gpt-oss-120b is our most powerful open-weight model, which fits into a single H100 GPU (117B parameters with 5.1B active parameters). Features permissive Apache 2.0 license, configurable reasoning effort (low, medium, high), full chain-of-thought access, fine-tunable parameters, and agentic capabilities including function calling, web browsing, Python code execution, and structured outputs.
Input: Text
Output: Text
Providers
deepinfra
Credits
Context131k
Max Output131k
Input$0.040/1M
Output$0.160/1M
Cache Read—
Cache Write—
groq
Credits
Context131k
Max Output131k
Input$0.150/1M
Output$0.750/1M
Cache Read—
Cache Write—
cerebras
Credits
Context131k
Max Output131k
Input$0.350/1M
Output$0.750/1M
Cache Read—
Cache Write—
openrouter
Credits
Context131k
Max Output131k
Input (Max)$0.370/1M
Output (Max)$0.790/1M
Cache Read—
Cache Write—
Quick Start
Use OpenAI GPT-OSS 120b through Helicone's AI Gateway with automatic logging and monitoring.