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

# Agents & Models

> The four agent providers, their model catalogs, reasoning-effort options, and interaction modes.

The agent is the AI coding system that runs inside your [workspace](/concepts/workspaces). Aether supports four selectable **agent providers**, each wrapping a real coding-agent CLI with its own curated model catalog and authentication. You choose the agent type, model, interaction mode, and reasoning effort per task — and as project or account defaults.

The platform defaults are agent type `codex`, interaction mode `default`, and the `gpt-5.6-sol` model.

## Agent Providers

### Codex (default)

OpenAI Codex. The default agent type. Authenticate with a ChatGPT subscription (device OAuth) or an API key.

| Model slug              | Name                |
| ----------------------- | ------------------- |
| `gpt-5.6-sol` (default) | GPT-5.6 Sol         |
| `gpt-5.6-terra`         | GPT-5.6 Terra       |
| `gpt-5.6-luna`          | GPT-5.6 Luna        |
| `gpt-5.5`               | GPT-5.5             |
| `gpt-5.4`               | GPT-5.4             |
| `gpt-5.4-mini`          | GPT-5.4 Mini        |
| `gpt-5.3-codex-spark`   | GPT-5.3 Codex Spark |

### Claude Code

Anthropic Claude Code. Authenticate with a Claude setup token.

| Model slug                | Name              |
| ------------------------- | ----------------- |
| `claude-opus-5` (default) | Claude Opus 5     |
| `claude-fable-5`          | Claude Fable 5    |
| `claude-sonnet-5`         | Claude Sonnet 5   |
| `claude-sonnet-4-6`       | Claude Sonnet 4.6 |
| `claude-haiku-4-5`        | Claude Haiku 4.5  |

### OpenCode

The OpenCode agent, with a curated catalog of around 30 models routed through OpenRouter, OpenCode Zen, and Anthropic — including Claude, GPT, Gemini, DeepSeek, Mimo, Kimi, Qwen, and GLM variants. The default model is `opencode-claude-sonnet-4-6`. Authenticate by uploading an `auth.json` file; runtime model availability depends on that file's provider entitlements.

### Cursor

The Cursor agent (the `cursor-agent` CLI). Authenticate with a Cursor OAuth subscription.

| Model slug                | Name                      |
| ------------------------- | ------------------------- |
| `claude-sonnet` (default) | Claude Sonnet 4.6         |
| `auto`                    | Auto (Cursor's selection) |
| `claude-haiku`            | Claude Haiku 4.5          |
| `gpt-5`                   | GPT-5.5                   |
| `gemini-3-pro`            | Gemini 3.1 Pro            |
| `grok-4`                  | Grok 4.3                  |

For the connect/authentication steps for each provider, see [connecting agent providers](/guides/connect-agent-providers).

## Reasoning Effort

Reasoning effort controls how much the model thinks before acting. The available options depend on the agent provider:

| Provider    | Default | Options                                                            |
| ----------- | ------- | ------------------------------------------------------------------ |
| Codex       | `xhigh` | `low`, `medium`, `high`, `xhigh`, `max`, `ultra` (model-dependent) |
| Claude Code | `xhigh` | `low`, `medium`, `high`, `xhigh`, `max`, `ultrathink`, `ultracode` |
| OpenCode    | `high`  | `low`, `medium`, `high`, `max`                                     |
| Cursor      | `none`  | `none`                                                             |

For Claude Code, the full option set above applies to Fable 5, Opus 5, and Sonnet 5; Sonnet 4.6 uses the adaptive set `low`, `medium`, `high`, `ultrathink`. Cursor does not expose reasoning effort — its thinking parameters are encoded into the model itself.

## Interaction Modes

A task runs in one of two interaction modes:

* **`default`** — the agent works the task end to end.
* **`plan`** — the agent drafts a plan and halts for your approval before making changes. You review the plan and approve or send feedback before the agent proceeds.

## Selecting Per Task

Pick the agent, model, mode, and effort when creating a task, or change them on an existing task:

```bash theme={null}
aether task create --project <project-id> --prompt "..." \
  --agent-type claude-code \
  --model claude-opus-5 \
  --interaction-mode plan \
  --reasoning-effort high
```

```bash theme={null}
aether task update <task-id> --agent-type codex --model gpt-5.5
```

You can also set these as project or account defaults so new tasks inherit them. See the [tasks guide](/guides/tasks) and the [agents and models guide](/guides/agents-and-models).