The model name isn’t matching what I set it up to be in the LLM settings. For example, it should be “Athena’s model” and not “GPT o4-mini-high” as seen in the screenshot below. I am indeed currently using the o4-mini-2025-04-16 model but this shouldn’t be displayed since I set up a specific name for it. Besides, showing which model is being used can be an attack vector for prompt injection since some models have specific weaknesses/vulnerabilities.
I just noticed after switching my instance over to the new Horizon theme, but I’m not sure if it’s directly correlated. I switched back to the default theme but the issue is the same.
. I only bring this up because it was working before as you described