For these release notes, we used an LLM (OpenAI gpt-5-mini) to review and summarise the code changes, along with the associated issue and PR descriptions. It may contain typical errors and inaccuracies. You can read the prompt here.