StepFun: Step 3.7 Flash

Step 3.7 Flash is a StepFun model that accepts text, image, and video input and returns text. It offers a 262,144-token context window with completions of up to 230,400 tokens, so long documents and lengthy outputs fit in a single request. It supports tool calling and reasoning; structured output support is not documented, so teams that depend on strict schemas should test it before committing. At $0.20 per million input tokens and $1.15 per million output tokens, it sits at the low end of pricing, which suits high-volume workloads, especially those that read heavily and write less. Its blended benchmark score of 29.1 comes from a single independent benchmark, where it scored 32.1 on the AI Index. One data point says little about overall quality, so treat its standing as unproven. Shortlist Step 3.7 Flash if you need cheap multimodal input with long context, and run your own evaluations before relying on it.