o3 Mini (batch)

A cost-efficient reasoning model from OpenAI with extended chain-of-thought processing optimized for STEM tasks, supporting configurable reasoning depth via the reasoning_effort parameter.

Strengths

Excels at science, mathematics, and coding problems while maintaining lower computational cost than larger reasoning models.

Best for

Batch processing of STEM reasoning tasks where extended chain-of-thought processing improves accuracy without the cost of o1-pro.

Limitations

The 200k token context is smaller than the 1 million token window available in the GPT-4.1 series, and reasoning capability is lighter-weight than o1-pro.