> ## Documentation Index
> Fetch the complete documentation index at: https://docs.pre.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Pro Mode & Model Selection

> Pin every phase to the strongest model, or pick models per phase.

Pro Mode is the quality dial turned all the way up: toggling `/pro` pins **every phase** of the agent's work — chat, research, coding, and acceptance — to the Pro model, **Claude Opus 4.8**.

## What it changes

Every sprint runs in phases, and each phase normally uses the default model (GLM 5.2) unless you've overridden it. Pro Mode overrides all of them at once:

|            | Standard                   | Pro Mode        |
| ---------- | -------------------------- | --------------- |
| Chat       | GLM 5.2 (or your override) | Claude Opus 4.8 |
| Research   | GLM 5.2 (or your override) | Claude Opus 4.8 |
| Coding     | GLM 5.2 (or your override) | Claude Opus 4.8 |
| Acceptance | GLM 5.2 (or your override) | Claude Opus 4.8 |

Toggle it off with `/pro` again and your previous per-phase settings apply.

## Credits

Pro Mode uses **more credits** per sprint — the Pro model costs more per token than the default. Check your remaining balance anytime with `/balance`, and see [pricing](https://pre.dev/pricing) for plan differences.

<Tip>
  A common workflow: iterate in standard mode while shaping a feature, then flip on Pro Mode for the final production pass. You can toggle between sprints without losing any project state.
</Tip>

## Per-phase control with /model

If all-Opus is more than you need, `/model` gives you the same power with precision — pin a strong model to just the phase that matters:

```
/model coding sonnet-4.6
/model acceptance opus-4.8
```

The full catalog:

| Model                    | Notes                           |
| ------------------------ | ------------------------------- |
| **GLM 5.2**              | The default                     |
| **Kimi K2.6** / **K2.7** | Strong open-weight alternatives |
| **MiniMax M3**           | Open-weight alternative         |
| **Claude Sonnet 4.6**    | High quality, balanced cost     |
| **Claude Opus 4.8**      | The Pro model                   |
| **GPT 5.5**              | OpenAI's flagship               |

Set a phase back to **Default** to return it to GLM 5.2.

<Note>
  For the full picture — sprint modes, effort levels, and how the model picker fits in — see [Build Modes & Effort](/coding-agent/building/build-modes).
</Note>