As a Goose maintainer, I know how much work goes into giving users a choice of models. We support more than 80 model providers, and keeping those integrations working is a lot of work. It's a drag managing the differences in tool calls, streaming, and model behavior across dozens of providers. Thankfully, Open Responses standardizes the LLM API format, so AI developers can write and maintain one integration instead of a separate adapter for each one.…