Choose the right AI model for each task in BrainBox
How to choose among more than 50 AI models in BrainBox, or let Auto choose for you, without losing your citations.
Lesson 11 of 21 · 1:47
What it covers
In BrainBox you choose among more than 50 AI models from Anthropic, OpenAI, Google, xAI, DeepSeek and others, or let Auto pick the best one for each question. Here you see how to read each model's usage, context and capabilities, and how the answer keeps its citations with any of them.
Frequently asked questions
›Which AI models can I use in BrainBox?
More than 50 models from labs like Anthropic (Claude), OpenAI (GPT), Google (Gemini), xAI (Grok), DeepSeek and Qwen, from the picker next to the microphone.
›What does the Auto option do?
It picks the best and most cost-efficient model for each question, so you don't have to decide every time.
›Do citations work the same with every model?
Yes. You change the model, but the answer still comes from your documents, with its citations.
Transcript
0:00The question
Julián, from IT, asks: which AI should we use, Claude, GPT or Gemini? The short answer: all of them. In BrainBox you choose.
0:12All the models
Claude, GPT, Gemini, Grok, DeepSeek, Qwen... more than fifty models, in one place. You choose which one to talk to, or let BrainBox pick.
0:26The picker
Open the picker, next to the microphone. Auto picks the best model, and the most efficient one, for each question. You can also filter by lab: Anthropic, OpenAI, xAI, Google. Each model tells you how much it uses, how much context it reads, and whether it sees images or reasons. Choose Claude Opus, and ask the usual question: what penalty do we charge Ferrolumina if it delivers late? Notice: the model changed, but the documents are the same. And the answer comes with its citations, as always.
1:06How to choose
How do you choose? For everyday work, Auto. If your documents are huge, look for more context. For hard analysis, a model that reasons. And watch the usage: the most powerful ones use more per question.
1:23Recap
To recap: you have more than fifty models in one place, Auto picks the best one for each question, and you change models without losing your citations. So, Julián: you don't have to marry a single AI. Change the model, not the documents.