Why is GPT-5.6 failing long-context tasks?

Most likely the Chat version is quantized however it is still entirely capable for regular chats and research. Only people who might suffer are those who use it for deep project analysis/planning instead of Work/Codex and especially those who attempt to use it to implement