New coding models & integrations · Ollama Blog
Key Takeaways
- GLM-4.6 and Qwen3-coder-480B are now available on Ollama's cloud service.
- Qwen3-Coder-30B has been updated for improved speed and reliability in tool calling.
- The article provides setup guides for integrating these models with VS Code, Zed, and Droid.
- An example prompt demonstrates model capability by requesting the creation of a complex HTML game application.
- Cloud models can also be accessed directly through Ollama's cloud API using an API key.
Ollama has announced the availability of GLM-4.6 and Qwen3-coder-480B on its cloud service, making them easily accessible for users, while simultaneously updating Qwen3-Coder-30B to feature faster and more reliable tool calling capabilities within Ollama's updated engine. The announcement includes direct command-line instructions for users to pull and run these models, including a local option for Qwen3-Coder-480B for those possessing over 300GB of VRAM. To illustrate the models' utility, an example prompt is provided requesting the creation of a feature-rich, colorful single-page game application named "Ollama's Adventure." The article extensively details integration steps for developers, covering setup procedures for using these models within VS Code, the newly Windows-compatible Zed editor, and the Droid CLI tool. Finally, it outlines how to access these cloud models directly via Ollama's cloud API using an environment variable for authentication.
