Model tiers

Every agent surface in Nexrall — Fullstack Agent Coding, Nexrall Code, the Nexrall CLI, Canvas — lets you pick a model tier. The tier controls which underlying model powers the agent loop for that session.

The tiers

Tier	Description
Turbo	Fast and the default choice — a large context window, tuned for everyday coding and chat. Best default for most tasks.
Pro	The most capable reasoning tier, for complex, multi-step problems where Turbo's answers aren't quite deep enough.
Ultra	Maximum power — a very large context window with adaptive deep thinking always on, for the hardest problems (large-codebase refactors, subtle bugs, long multi-file plans).

Choosing a tier

• Start with Turbo for most day-to-day work — it's fast and handles the large majority of coding and automation tasks well.
• Reach for Pro when a task needs deeper reasoning: unusual architecture decisions, tricky debugging, or a plan spanning many steps.
• Reach for Ultra for the hardest, highest-stakes tasks, or when you're working across a very large codebase and need the largest context window.

Where to switch tiers

• Web app — the model picker in the chat composer.
• VS Code — the model menu in the chat panel header.
• CLI — /model inside a session, or --model <tier> / the shorthand flags (-u for Ultra) on the command line.

Context window

All three tiers support very large context windows, which matters for two things:

• Long agent sessions — many turns of back-and-forth without losing earlier context.
• Large codebases — reading enough of a project at once to make a correctly-scoped change instead of a locally-correct-but-globally-wrong one.

When a session's context does approach the limit, Nexrall automatically compacts older history — summarizing it while pinning your original request and a structured progress ledger verbatim, so the agent doesn't drift from the original goal even after several rounds of compaction.