GLM-4.6 and Qwen3-Coder models available on Ollama

Ollama announces GLM-4.6 and Qwen3-Coder-480B on its cloud service, with a new engine update for Qwen3-Coder-30B tool calling.

By OpenSmartRoute editorial · written through the router by writer-small

From Ollama blog - “New coding models & integrations”

This post is in the archive: it was published on Oct 16, 2025 and the models, prices or APIs it describes may have changed since. See the latest posts.

GLM-4.6 and Qwen3-Coder models available on Ollama — Image: Ollama blog (original)

Ollama has made GLM-4.6 and Qwen3-Coder-480B available on its cloud service. These models offer easy integrations with familiar tools for developers managing AI systems.

Qwen3-Coder-30B received an update to improve the speed and reliability of tool calling within Ollama's new engine. For users possessing more than 300GB of VRAM, the Qwen3-Coder-480B model is also available for local deployment.

Integration with Coding Tools

Documentation now includes sections on using Ollama with popular coding tools. Models can be accessed directly via VS Code and Zed through the provider dropdown interface.

The cloud API allows direct access to models like GLM-4.6 and Qwen3-Coder-480B. Users must create an API key and set it in their environment, configuring the base URL and provider settings in their application configuration files.

Source: https://ollama.com/blog/coding-models

Keep reading

TensorRT Edge-LLM Runs Qwen3.6 on Jetson AGX Thor

TensorRT Edge-LLM ran Qwen3.6-27B on a Jetson AGX Thor. It finished the MLPerf Edge Agentic benchmark 6.4 times faster than the reference run.

LLMsSep 16, 20261 min read

Ten audit tools detect bias but fail to rank frontier models consistently

A study of ten frontier models using ten extrinsic audit instruments found that while all tools detected bias, their rankings were statistically indistinguishable from chance.

GLM-4.6 and Qwen3-Coder models available on Ollama

Integration with Coding Tools

Related posts

TensorRT Edge-LLM Runs Qwen3.6 on Jetson AGX Thor

Ten audit tools detect bias but fail to rank frontier models consistently

MIMIC: Code-Centric Data Synthesis for LLM Reasoning