In practice
Through an app connected to the model, an interview editor could supply a one-minute clip and interview outline, request key points, quotes and a suggested edit order, then check each against the footage.
Facts: access and cost
The model documentation lists Beijing, Singapore and other regions, accessed through Chat Completions or Responses with a regional API key. On September 19, Beijing lists CNY 0.8 per million input tokens and CNY 2.7 per million output tokens, with separate cache-hit pricing. Other regions differ; these are not per-minute video rates.
Editorial assessment: start with material review
The useful checks are who said what and what the footage actually shows. Keep quotes, visual descriptions and editing suggestions separate. Misattribution, omitted conditions or invented scenes need correction against the source.
Limits and verification
We checked official releases and documentation; we have not tested recognition accuracy, editing integrations or completion time. Tool calls require external execution. Native multimodal input does not mean automatic video editing, generated voice output or identical availability in every Qwen chat account.