Over the past few month, we’ve noticed more developers pushing the limits of TRAE: working with larger repositories, drafting long design documents, and chaining together complex, multi-step reasoning tasks.
With TRAE Pro, you have a solid foundation—adequate context windows and tool calls for most day-to-day work. But when you're building a multi-service application or or a codebase that stretches across many files, the limits start to show.
So we are introducing Max Mode, which is designed for deep coding with larger context windows and more tool calls.
What Makes Max Mode Different
-
200k context window. Big enough to hold entire codebases, lengthy specs, and deep technical context, so the agent can keep everything in view without losing coherence or being cut short.
-
Up to 200 tool calls. With this expanded limit, the agent can plan, execute, test, and refine in longer workflows. There's no need to keep hitting “continue” to get your work done.
-
Claude-4-Sonnet Beta, Claude-3.7-Sonnet, and Claude-3.5-Sonnet are fully supported.
This makes Max Mode especially useful for:
-
Losing context mid-session breaks your flow.
-
You need to run long chains of builds or tests.
-
You would like the agent to “run long” without intervention.
Understanding Max Mode billing and usage details
If you're already a Pro user, you can choose to switch on Max Mode inside your TRAE IDE. Your Pro plan remains unchanged, and it still includes 600 Fast Requests for your regular workflows.
Here’s how billing works when Max Mode is activated:
-
Token consumption is tracked per session—including input, model output, and cache reads/writes.
-
The total tokens are multiplied by the model’s per-token rate to calculate a cost.
-
That cost is converted into Fast Request equivalents, which are then deducted from your remaining Fast Request balance - from your subscription or any existing extra packages you purchased. Max Mode token consumption will be deducted from whichever expires first.
For example, say you use 1,000 tokens in a Max Mode session with a token rate of $0.001. That’s $1 total, which converts to roughly 58 Fast Requests (assuming $0.017 per Fast Request). If you started with 600, you’d now have 542 left.
You can find your Max Mode usage (token, cost, Fast Request) within IDE after every round of conversation or you can find all your historical usage under Usage in your account profile.
In summary, for small tasks, the current fast request in Pro is still a great option. But for more complex jobs that require deep coding, Max Mode lets you break past with larger context windows and more tool calls.