OpenAI: GPT-5 Mini

GPT-5 Mini is OpenAI's smaller-tier text and vision model, accepting text, image, and file inputs with a 400,000-token context window and up to 128,000 tokens of output per call. It supports tool use and reasoning, which makes it usable for agentic workflows and multi-step tasks. Structured output support is unconfirmed in available specs, so teams with strict schema requirements should verify that before committing. At $0.25 per million input tokens and $2.00 per million output tokens, it sits in the budget tier of capable multimodal models. Its blended benchmark score of 54.2 comes from a single benchmark, so that figure gives limited confidence about broad performance. Buyers who prioritize low input cost and need a long context window for document-heavy or agentic tasks have reason to shortlist it, but those who need well-rounded performance validation should weigh the thin benchmark coverage carefully.