OpenAI: o4 Mini High

OpenAI o4 Mini High is a paid model from OpenAI that accepts text, images, and files as input, with a 200,000-token context window and up to 100,000 tokens of completion per response. It supports tool use and reasoning, making it suitable for agentic workflows and multi-step tasks that benefit from structured problem-solving. On pricing, it sits at $1.1 per million input tokens and $4.4 per million output tokens, which positions it below many frontier-tier models in input cost while keeping output costs moderate. Its blended benchmark score of 81.0 comes from a single benchmark (aider_polyglot at 72.0), so that figure carries limited confidence across diverse task types. Developers focused on coding tasks who want reasoning capabilities at a mid-range price may find it worth shortlisting, but buyers needing broad, verified performance should treat the benchmark coverage as thin until more independent results are available.