From e7ac9b4563c3e23cf47226b29c720b08c56386ad Mon Sep 17 00:00:00 2001 From: Thomas Chopitea Date: Tue, 1 Sep 2026 14:35:52 +0000 Subject: [PATCH] Offer a list of models in the agents example configuration GEMINI_MODEL named one model. The chat can offer a choice now, so the example lists them instead, and the first is what a conversation uses until one is picked for it. The id it suggested was gemini-3-flash-preview, which is a preview and was already behind; the list names stable ids, with the one preview marked as such because those are retired on Google's schedule rather than ours. --- prod/.env.example | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/prod/.env.example b/prod/.env.example index 4939e26..8d2bf98 100644 --- a/prod/.env.example +++ b/prod/.env.example @@ -42,10 +42,14 @@ YETI_AGENTS_WEBSOCKET_ROOT=ws://agents:8888 # Gemini configs # LLM_PROVIDER=gemini # GOOGLE_API_KEY=API_KEY -# GEMINI_MODEL=gemini-3-flash-preview +# Models the chat offers, best first. The first is what a conversation uses +# until one is picked for it, and the selector is hidden when only one is +# listed. Keep this to what you are willing to spend quota on rather than +# everything Google publishes; -preview ids are retired on their schedule. +# GEMINI_MODELS=gemini-3.7-flash,gemini-3.5-flash,gemini-3.5-flash-lite,gemini-3.1-pro-preview # # Ollama configs # LLM_PROVIDER=ollama # OLLAMA_API_BASE=http://ollama:11434 # Pick a model supporting tools and thinking: https://ollama.com/search?c=thinking&c=tools -# OLLAMA_MODEL=gemma +# OLLAMA_MODELS=gemma