Models

Tyrex AI runs three working models with different depth. They're not marketing tiers — they're genuinely different tools, and you can switch between them per chat.

RAPTOR — the everyday brain

• Speed: Fast
• Availability: All plans
• Usage: Most efficient units

RAPTOR is your daily driver: quick, light, and always ready. It handles routine work brilliantly — chat, Q&A, answering customers, quick summaries, and tasks that don't need deep thinking.

Best for: chat & Q&A, answering customers, quick summaries, light tasks.

When to use it: most conversations. If a task feels routine, RAPTOR is the efficient choice — you'll get fast answers and stretch your quota further.

CARNO — the reasoning brain

• Speed: Balanced
• Availability: Pro & Elite
• Usage: Medium units

CARNO thinks harder. It's built for work that needs real reasoning: analyzing data, researching a topic thoroughly, working through complex documents, and producing more mature, carefully considered results.

Best for: data analysis, deep research, complex documents, heavy tasks.

When to use it: when the answer matters more than the speed. Switch to CARNO for analysis, long-form writing, or anything where RAPTOR's quick take isn't enough.

REX — the flagship brain

• Speed: Deep
• Availability: Elite
• Usage: Highest units

REX is our most powerful model. It's for the most demanding work: big projects, premium long-form content, and the absolute highest quality output — when the outcome is important enough to spend the units.

Best for: big projects, premium content, heaviest tasks, best output.

When to use it: flagship work. If you're producing something that will be seen, shared, or sold — REX is the difference.

Choosing per chat

Models are switched per conversation, not per account. A single session can start on RAPTOR for quick warm-up questions, switch to CARNO for the deep analysis, and finish on REX for the polished deliverable. Context carries across switches.

In the app: the model switcher sits at the top of any chat (Chat in the sidebar). You can also set a default model when creating a persona, so every new chat with that AI starts on the brain you chose.

Which models your plan unlocks is visible on the Plans page, and the switcher only offers models you can actually use.

The usage meter makes the tradeoff visible: lighter models are more efficient, heavier models consume more units. You always see the cost of the brain you picked.

On the horizon

Two more models are in active development, reserved for the biggest jobs and future tiers:

• TITAN — built for the largest workloads: massive context, extended reasoning, and tasks where even REX reaches its limits.
• GIGA — the next step in raw capability, for the heaviest enterprise-grade work.

Both will arrive with clear availability, pricing, and use cases — the same transparency as the models above.

Next: API & Webhooks.