Guided Model Selection

With so many models available, choosing the right one can be overwhelming. Guided model selection helps you pick a model even if you are not familiar with technical details.

How it works

The model selector offers two modes. By default, guided selection walks you through a performance-based choice:

1. Choose a performance tier — Three tiers appear as cards: Fast & Efficient (quick, cost-effective for everyday use), Balanced Performance (optimal speed and quality for most tasks), and Maximum Quality (best for complex reasoning and demanding tasks). Each card shows how many models are available in that tier and the pricing range ($/M tokens for read and write).
2. Pick a model — After selecting a tier, you see up to five models. Each card shows the lab name (who created the model), optionally via the serving provider when it differs, the model name, optional description, and per-million-token pricing, so you can compare quickly.

Only models available on your configured provider(s) appear. You do not need to know model IDs or provider APIs; the interface guides you to a suitable model.

Pro Mode — Experts can toggle Pro Mode in the model selector to bypass tiers and choose freely from the complete list of models. Pro Mode shows a searchable, sortable table with full details (including model metadata such as lab, provider, country of origin, and capabilities). Your Pro Mode preference is remembered across sessions.

Switching models

You can change the selected model at any time during a conversation. The next message will use the new model. This lets you try different models for different tasks without starting a new chat.

Tips

• For general chat, any of the recommended or default models is a good starting point.
• For image generation, pick a model that supports the image generation tool.
• If you care where the model comes from or where data is processed, use model metadata (lab and provider country of origin) and Data routing policy to narrow your choice.