Hi, Hume team! I hope you're well. Is there any way you could directly support ChatGPT-4o-latest for EVI 2? The only way to currently use that supplemental LLM is to set up a CLM which, as mentioned in your docs, increases latency.