Changing models

Override the default model for a provider on a single run or permanently through configuration.

Every built-in provider ships with a default model. You can change it in two ways.

One-time override

Pass --model directly when launching. The flag is forwarded to Claude CLI.

clother-zai --model glm-4.7
clother-alibaba --model kimi-k2.5
clother-ollama --model qwen3-coder

The override applies only to that single session. The configured default is unchanged.

Permanent override

Run clother config <provider> and choose a different default model. Clother writes the override to config.json and uses it for all future launches of that provider.

clother config zai
# Choose model:
#   1. glm-5     GLM-5
#   2. glm-4.7   GLM-4.7
# Model [glm-5]: 2

To see what model is currently in effect for a provider:

clother info zai

Provider default models

Provider	Launcher	Default model	Other available models
Anthropic (native)	clother-native	(Claude default)	—
Z.AI	clother-zai	glm-5	glm-4.7
Z.AI China	clother-zai-cn	glm-5	glm-4.7
MiniMax	clother-minimax	MiniMax-M2.5	MiniMax-M2.5-highspeed, MiniMax-M2.1, MiniMax-M2.1-highspeed, MiniMax-M2
MiniMax China	clother-minimax-cn	MiniMax-M2.5	MiniMax-M2.5-highspeed, MiniMax-M2.1, MiniMax-M2.1-highspeed, MiniMax-M2
Kimi	clother-kimi	kimi-k2.5	—
Moonshot AI	clother-moonshot	kimi-k2.5	—
DeepSeek	clother-deepseek	deepseek-chat	—
Xiaomi MiMo	clother-mimo	mimo-v2-flash	—
Alibaba (Singapore)	clother-alibaba	qwen3.5-plus	kimi-k2.5, glm-5, MiniMax-M2.5, qwen3-coder-next, qwen3-coder-plus, qwen3-max-2026-01-23, glm-4.7
Alibaba (US)	clother-alibaba-us	qwen3.5-plus	kimi-k2.5, glm-5, MiniMax-M2.5, qwen3-coder-next, qwen3-coder-plus, qwen3-max-2026-01-23, glm-4.7
Alibaba China	clother-alibaba-cn	qwen3.5-plus	kimi-k2.5, glm-5, MiniMax-M2.5, qwen3-coder-next, qwen3-coder-plus, qwen3-max-2026-01-23, glm-4.7
VolcEngine	clother-ve	doubao-seed-code-preview-latest	—
Ollama	clother-ollama	(specify via --model)	any model pulled locally
LM Studio	clother-lmstudio	(specify via --model)	any model loaded in LM Studio
llama.cpp	clother-llamacpp	(specify via --model)	any model served by llama-server

OpenRouter

OpenRouter launchers follow the clother-or-<alias> naming pattern. The model is bound to the alias at configuration time, not at launch time.

1. Run the OpenRouter configurator

   clother config openrouter
   
2. Enter a model ID and alias

   Model ID (empty to stop): moonshotai/kimi-k2.5
   Alias [kimi-k2-5]: kimi-k25
   

   Find model IDs at openrouter.ai/models.

3. Use the launcher

   clother-or-kimi-k25
   

To use a different model through the same alias, re-run clother config openrouter and provide a new model ID for the same alias name.

If a model doesn’t behave as expected with tool calls, try the :exacto variant — for example moonshotai/kimi-k2-0905:exacto.