How AI models are selected and updated

Goal

Understand which AI models BoodleBox offers, how new ones are added, and what happens when one is retired.

Many models, many providers

BoodleBox provides access to leading models from OpenAI, Anthropic, Google, Meta, Amazon, NVIDIA, and Mistral, plus web-research models (Perplexity, Semantic Scholar, Tavily) and image-generation models. Some are free (Boodle Basic) and some are premium (Boodle Premium).

How new models are added

• A 4-week deployment schedule behind the provider's release. When a provider ships a new model, BoodleBox waits about four weeks before adding it. New models sometimes ship with bugs; the buffer confirms there are no major security issues before exposing it to customers — important for education.
• Two generations stay available. When a new version lands, the prior generation stays too, so nothing changes under you overnight.

What happens when a model is retired

• Your bots don't break. When a model is deprecated, custom bots pointing at it are automatically migrated to the closest current equivalent.
• During the overlap window you can test the newer model on your own bots before the auto-migration.

How auto-routing chooses

When you use BoodleBot, it routes each message to the best model for the task — based on what you're trying to do, not on cost. (See How BoodleBot picks the right AI for you.)

Questions or documentation?

Governance and security docs are on the Trust Center: https://trust.boodlebox.ai/. For specifics, email compliance@boodle.ai. General help: success@boodle.ai.