| Harness | CLI | Default model | Effort |
|---|---|---|---|
claude | Claude Code | opus | low medium high xhigh max |
codex | Codex | gpt-5.5 | low medium high xhigh |
opencode | OpenCode | glm-5.2 | default minimal low medium high max |
sov | Sovereign AI | alias set | off low medium high max |
grok | Grok Build | grok-4.6 | depends on model (max is not on grok-4.5) |
sonnet opus fable haiku.
Codex: gpt-5.5 gpt-5.5-codex gpt-5.6 gpt-5.6-sol gpt-5.6-terra gpt-5.6-luna (extend with TELEKIT_CODEX_MODELS).
OpenCode: glm-5.2 glm-5v-turbo glm-5.1 glm-5-turbo (extend with TELEKIT_OPENCODE_MODELS).
Grok: grok-4.6 grok-4.5 4.6 4.5 grok.
SOV aliases: sonnet opus haiku openai-gpt openrouter-sonnet.
/effort short tokens via /ef: l m h xh xx. /effort auto is valid only while /ack is on.
Setting the same effort already in force does not rewrite.
Provider (SOV and OpenCode)
Refused on Claude, Codex, Grok. Native SOV labs:openrouter openai anthropic. Other labs ride OpenRouter with that lab’s default model. A set provider wins over the lab baked into a /model alias. /provider off clears it.
/fanout
The rest of the line is the prompt. The harness splits work onto native subagents.
Supported: Claude, Codex, Grok, OpenCode. Not SOV.
Needs a prompt. Refused if a turn is already running. Plan-gated on some credits tiers.
Pin binaries with TELEKIT_CLAUDE_BIN, TELEKIT_CODEX_BIN, TELEKIT_OPENCODE_BIN, TELEKIT_SOV_BIN, TELEKIT_GROK_BIN (also pinnable into the LaunchAgent at telekit install).