Models and Bloxks Empire
Auto picks the right AI for each task. Pro and Empire let you choose your model.
Auto: the default that just works
By default, Bloxks runs in Auto mode: it picks the best model for each request. Planning a new app, building screens, fixing a detail: each task goes to the model that handles it best, at the best speed. Most of the time, you never need to think about it.
Choose your model on Pro and Empire
On the Pro and Empire plans, the Auto pill becomes a real model picker. You can pin every request to a specific model. This lineup is updated regularly as better models become available; here is the current one:
- Bloxks v1 ($): Fast, capable, best value
- Claude Fable 5.1 ($$$$): Anthropic model for complex coding and reasoning
- GPT-6 Astra ($$$$): OpenAI's most capable model for the hardest end-to-end work
- GPT-6.1 Sol ($$): OpenAI's balanced model, near-Astra quality at a lower price
- GPT-6.1 Terra ($$): OpenAI's fast mid-tier model for everyday building
- Grok 4.7 ($$): xAI's coding model, strong reasoning at a low price
- Claude Opus 5.5 ($$$): Flagship for the hardest builds
- Claude Sonnet 5.5 ($$): Balanced — strong and quick
They do not all cost the same
This is the part most people miss, and it is the one that weighs most on your balance. Hover any model in the picker and its cost shows up next to its name, from $ (the cheapest) to $$$$ (the most expensive):
The gap is not small: a $$$$ request can cost around thirty times a $ one.
A more powerful model does not give a better result on a small task. Changing a label, fixing a button, adjusting a colour: a $ model does that just as well, for a fraction of the price. Keep the expensive models for what actually needs them: the first version of an app, a whole feature, a bug that resists.
The Bloxks Empire toggle
Empire subscribers also get the Bloxks Empire toggle. Switch it on and Bloxks takes extra time to think before building: it plans deeper, weighs options and produces noticeably more polished results. Use it for the moments that matter: the first version of your app, a big redesign, a feature you want perfect.
When to pick what
| You are doing | Good pick | Cost |
|---|---|---|
| A quick text or color change | Bloxks v1 | $ |
| Everyday features | Auto, or Sonnet 5.5 | $$ |
| A bug that resists | Opus 5.5 | $$$ |
| Starting a new app | Bloxks Empire on, or Fable 5.1 | $$$$ |
Good to know
- Switching models never breaks your project: they all work on the same app.
- Model choice is per message: you can build with Sonnet 5.5 and polish one screen with Fable 5.1 right after. It is also billed per message, so leaving the picker on an expensive model costs you on every single request, including the tiny ones.
- In doubt, leave Auto on: it picks a model sized for the task instead of always taking the most expensive one.