Auto
Uses Flash for everyday and photo work, and Gemini for questions identified as needing research or current information.
Loading your account…
Dino Chat is getting ready. Explore the site — accounts, chat, and subscriptions are not open yet.
Choose your helper
Start with Auto. Change the model when you have a reason to try a different approach; no model is best at every task.
| Choice | Useful for | Photo questions |
|---|---|---|
| Auto | A simple starting point for most tasks | Yes, with a compatible model |
| Flash | Everyday questions, writing, and photos | Yes |
| Gemini | Research and photo-based questions | Yes |
| Pro | Another option for text problems | No; choose Auto, Flash, or Gemini |
Uses Flash for everyday and photo work, and Gemini for questions identified as needing research or current information.
A manual model choice for drafting, explaining, planning, and photo questions. You can still explicitly ask for a lookup when search is available.
A manual model choice for research and visual work. Choosing it does not itself run a web search or connect a Google account.
An explicit text-only alternative. Dino does not automatically switch to Pro just because a question is complicated.
Auto is a routing choice, not a fourth AI model. Current Flash and Pro choices use DeepSeek models; Gemini uses a Google model. Service availability and compatible input capabilities can limit what runs. An incompatible photo request should ask you to choose a suitable model rather than discard the photo.
For information that changes, ask “look this up” and request links to sources. Live search must be enabled and available. Gemini alone does not mean Dino checked the web, and it does not grant access to your Gmail, Drive, or Calendar.
Check cited pages, dates, and source quality. If lookup is unavailable, treat an answer as a starting point and verify it elsewhere. Google Calendar access requires a separate connection and supported changes require your approval.
Flash, Gemini, Pro, supporting AI calls, and enabled metered tools draw from the same allowance: $2.00 total during the 7-day trial, then $8 per monthly window after confirmed paid access. Costs vary with the model, input, output, and tools used. A longer or more involved answer can use more allowance; there is no fixed message count.
Pro is not a promise of better accuracy, and a more expensive model is not a guarantee of a better answer. Try a focused follow-up or supply better context before switching. Check your current usage in Billing.
Every model can miss context, invent details, or make mistakes. Verify important answers, especially decisions involving money, health, law, or safety.