Models and reasoning levels
HiCode, low, medium and high levels, and how to view usage.
Updated on Sep 16, 2026
HiNow Code works with HiCode, HINOW's code model: a 1 million token context window, built to read entire projects and work on long tasks.
Before acting, the model thinks, and you choose how much. In the interface, through the command palette (ctrl+p); in run, with --variant.
| Level | For what |
|---|---|
low | Objective tasks: rename, list, small fixes. Quick response |
medium | Day-to-day: features, local refactors, tests |
high | Architecture, difficult bugs, changes across the project |
It's the same control that the API exposes as reasoning.effort; the documentation is at Reasoning (thinking).
hinow-cli models # todos
hinow-cli models hinow # só os da HINOW
hinow-cli models --verbose # com custos e limiteshinow-cli stats shows tokens and cost by period, tool, and model. Consumption counts toward your plan's daily quota; see plans and pricing.
hinow-cli stats --days 7 --modelsWas this page helpful?

