🔄 We just refreshed the AI models in Fabio AI Chatbot — curious about your experience with them

We just went through and updated all the AI providers available in Fabio AI Chatbot (WordPress AI chatbot plugin) to keep up with the latest releases:

OpenAI — swapped out gpt-5.4, gpt-5.5, and gpt-5.4-mini for gpt-5.6-luna, keeping gpt-5.4-nano around too.

Google Gemini — moved to gemini-3.5-flash and gemini-3.1-flash-lite, dropping the discontinued/outdated previews. (Also worth flagging: gemini-2.5-flash-lite is being deprecated by Google around October 2026, if you're still on it elsewhere.)

Mistral — simplified down to just mistral-large-3.

Anthropic — brand new provider support, with claude-haiku-4-5 and claude-sonnet-5.

Now the part we're actually curious about: have you used any of these newer models in production? We'd love to hear real experience, not just benchmarks — things like:

• Response quality vs. cost trade-offs you've noticed

• Latency in real-world (not sandbox) conditions

• Any quirks or gotchas migrating off older versions

• Which one you'd default to for a customer-facing chatbot specifically

We picked defaults based on what looked strongest on paper, but nothing beats hearing what's actually working for people running these day to day. Drop your experience below 👇