← Chat Archive

To change the model in Ollama, you have a few options depending on what you mean by "change": ## Pull a different model ```bash ollama pull llama3.2 ollama pull mistral ollama pull phi4 ``` ## Run a specific model interactively ```bash ollama run llama3.2 ``` ## Switch models in a running `ollama run` session If you're already in an interactive chat, exit with `/bye` or **Ctrl+D**, then start with the new model: ```bash ollama run <model-name> ``` ## List available models ```bash ollama list ``` ## Set a default model (via aliases) Ollama doesn't have a global "default model" setting, but you can create an alias: ```bash ollama cp llama3.2 default ollama run default ``` ## In code / API calls If using the Ollama API, specify the model in your request: ```bash curl http://localhost:11434/api/generate -d '{ "model": "llama3.2", "prompt": "Hello" }' ``` What context are you changing the model in — CLI chat, API call, or a tool like OpenWebUI?

What task should I route? `/model-route` needs a task description. The current conversation is just a simple Q&A about Ollama commands, which is **haiku**-level. If you have a new task, describe it like: ```text /model-route refactor the auth cache logic in our OpenAlgo Flask app --budget med ``` Then I’ll recommend: | Tier | Best For | |------|----------| | **haiku** | Deterministic changes, docs, simple fixes | | **sonnet** | Implementation, refactors, feature work (default) | | **opus** | Architecture, deep review, ambiguous requirements |