Venice large 1.1 caution

I have not found the Venice 1.1 Large model to be as useful as claimed in the documentation. It tends to hallucinate more, even on topics that can be easily verified on Google (common physics answers or search terms for example.) When confronted with contrary information, I have found it hangs several times. It sounds more authoritative, but is actually less so. I would say that in terms of information searches, it seems less accurate and more “opinionated”,. meaning prone to hallucination than previously. I much prefer the older Llama 3.1 405B model, which is at least more comprehensive and accurate in its responses. It is more natural for conversational flow and remain focused on the task or query at hand without going off on irrelevant tangents.