Model and reasoning overrides

Untether lets you override which model the agent uses and its reasoning level, per chat or per engine — all from Telegram, without editing config files or restarting.

Check current model#

Send /model to see the override for the current engine and where it comes from:

/model

Untether

engine: claude (chat default)

model: sonnet (chat default)

defaults: topic: none, chat: sonnet

available engines: codex, claude, opencode, pi

model: **default** (no override) means no chat or topic override is set, so the engine uses its own default (see the resolution order below).

Set a model override#

Use /model set to override the model for the current engine:

/model set sonnet

To target a specific engine, include the engine name:

/model set claude opus

The override applies to the current chat (or topic, if you’re in a forum thread).

OpenCode: use provider/model format

OpenCode requires the provider/model format for model overrides (e.g. openai/gpt-5.5, anthropic/claude-sonnet-5-5). Using just the model name will fail. Example: /model set opencode openai/gpt-5.5.

Clear model override#

Remove the override to revert to the default:

/model clear

To clear the override for a specific engine:

/model clear claude

Set reasoning level#

Some engines support reasoning levels that control how much thinking the model does before responding. Use /reasoning set:

/reasoning set high

Valid levels depend on the engine:

• Claude Code: low, medium, high, xhigh, max (passed as --effort)
• Codex CLI: low, medium, high, xhigh

minimal was removed for Codex in v0.36.0: no current Codex model supports it, and it fails alongside Codex’s default web search. A saved minimal is ignored with a one-line note on each run, and the run uses the engine default until you pick another level.

Other engines (OpenCode, Pi, Gemini, Amp) ignore this setting.

Per-engine reasoning#

Target a specific engine with the engine name:

/reasoning set claude high

Clear reasoning#

Remove the reasoning override:

/reasoning clear

Or for a specific engine:

/reasoning clear claude

View full resolution#

Use /agent to see how all configuration layers resolve for the current scope:

/agent

The resolution order for model and reasoning is (highest priority first):

1. Cron override — a [[triggers.crons]] entry’s own model / reasoning, for that scheduled run only (#743; see Schedule tasks)
2. Topic override — set via /model set or /reasoning set in a forum topic
3. Chat default — set via /model set or /reasoning set in a private or group chat
4. Engine config — model under [engines.claude], [engines.opencode] or [engines.pi] in untether.toml (Codex takes its model from its own ~/.codex/config.toml or profile; there is no untether.toml default for reasoning)
5. CLI default — whatever the engine CLI picks on its own

There is no per-project model setting; projects only pick a default_engine.

Quick check

/agent shows the effective engine, model, and reasoning for the current context, including which layer each setting comes from.

Admin-only in groups#

In group chats, model and reasoning changes require admin or creator status. This prevents non-admin members from switching to expensive models or changing settings that affect everyone in the group.

Related#

• Switch engines — change which engine handles messages
• Commands & directives — full command reference
• Configuration — config reference for model and reasoning settings