API rate limits constricted?

A user reported significant issues with OpenAI API rate limits when using Luna in Lumina, noting that it restricts users to either reasoning or tool calls. The user found tool calls to be severely throttled, often requiring multiple "continue" prompts and taking ten times longer for agentic tasks. This led to switching to GLM 5.3, which handled extensive coding tasks with numerous reasoning steps and tool calls without similar throttling, raising questions about OpenAI's paid service limitations.