ollama launch

Ollama announced a new command called ollama launch on January 23, 2026. According to the company, the command sets up and runs coding tools including Claude Code, OpenCode and Codex with local or cloud models, and requires no environment variables or config files.

Users are directed to download Ollama v0.15 or later, then run commands from a terminal. The documented examples include pulling glm-4.7-flash, which the announcement says requires roughly 23 GB of VRAM with a 64000 token context length, or pulling glm-4.7:cloud to use a cloud model with full context length. Setup commands listed include ollama launch claude and ollama launch opencode, which guide users to select models and launch the chosen integration.

Supported integrations named in the announcement are Claude Code, OpenCode, Codex and Droid.

Ollama lists recommended coding models in two groups. Local models are glm-4.7-flash, qwen3-coder and gpt-oss:20b. Cloud models are glm-4.7:cloud, minimax-m2.1:cloud, gpt-oss:120b-cloud and qwen3-coder:480b-cloud.

The announcement notes that coding tools work best with a full context length and advises updating the context length in Ollama's settings to at least 64000 tokens, pointing to context length documentation.

For extended coding sessions, Ollama says its cloud service offers hosted models with full context length and generous limits even at the free tier, and that the update provides more usage and an extended 5-hour coding session window. Pricing details are directed to ollama.com/pricing. A configure-only option is also documented: ollama launch opencode --config configures a tool without launching it immediately.