OpenAI’s o4-mini-deep-research model offers fast, cost-efficient research and tool use via the Responses API, with $2 per million input and $8 per million output tokens, plus fees for web searches and code interpreter sessions.

In practical testing, it synthesized data across multiple tools to locate surviving orchestrions worldwide, delivering structured results but incurring additional per-tool-call costs, totaling about $1.10 for a large query. Output accuracy was limited by source reliability, and results were not fully verified or comprehensive.