gpt-oss-120b (batch)

An open-weight 117-billion-parameter Mixture-of-Experts model from OpenAI with a 131k token context window.

Strengths

Delivers larger model capacity than gpt-oss-20b while maintaining efficient inference through sparse activation of its expert mixture.

Best for

Batch processing workloads where inference latency is not a constraint and model capability outweighs deployment speed requirements.

Limitations

As an open-weight model released in 2025, it is positioned below the newer GPT-5 family in capability, and the batch-only variant introduces latency tradeoffs unsuitable for real-time applications.