Choosing between GPT-6.2 Astra and Sol for an in-app assistant?

I’m building a desktop app with wxPython that includes a simple text helper panel, and I’m trying to decide which model to wire into it. The app is small, mostly used on Windows machines in a local office, and the assistant only needs to summarize short notes and rephrase selected text. I keep coming across mentions of GPT-6.2 Astra and Sol and the lower pricing being talked about online, but I can’t work out which one makes more sense for a lightweight background feature like this. My main worry is latency and rate limits rather than raw capability, since the app would call the API frequently and any delay would make the UI feel sluggish. I’m also unsure whether the cheaper option is good enough for short, well-scoped prompts or if it will need more retries. For anyone running a model call from a wxPython worker thread, which of these have you settled on, and what made the difference?