Z.AI GLM model integration

[robert-zaremba]

Recent release of GLM 4.6 shows outstanding agentic performance with very attracive prices.

• Integrate Z.AI provider directly (PS: bigmodel.cn is also a good provider for GLM).
• Support GLM settings.
• tip: would be cool to Zed into z.ai supported tools section.

[bradsjm]

I'm using their coding plan package and it is working excellently with Zed just by adding:

    "language_models": {
        "openai_compatible": {
            "Z.ai": {
                "api_url": "https://api.z.ai/api/coding/paas/v4",
                "available_models": [
                    {
                        "name": "glm-4.6",
                        "display_name": "GLM-4.6",
                        "max_tokens": 200000,
                        "max_output_tokens": 128000,
                        "max_completion_tokens": 128000,
                        "capabilities": {
                            "tools": true,
                            "images": false,
                            "parallel_tool_calls": true,
                            "prompt_cache_key": true
                        }
                    }
                ]
            }
       }
   }

[azadoss]

where did you change it?

[MRiabov]

where did you change it?

project settings, isn't it?

[azadoss]

Max Dobro. Ja uje razobralsja. Would be great if it can be implemented in Claude Code?

[MorePeanuts]

I agree. GLM's coding plan is very affordable and works quite well, worthy of being a built-in provider option.

[LucasMatuszewski]

Now you can also use Ollama Cloud for GML-4.6 and use Zed's Ollama integration:
https://ollama.com/library/glm-4.6

On Ollama (free account) I get ~30-45 t/s, so not stable, but it's quite ok, considering I don't pay for it ;)
But stable ~50 would be much more comfortable (for qwen3-coder:480b-cloud I get >90 t/s)

Does anybody have Ollama subscription (20 USD) to check if it will be faster?

Does anybody know the speed of Z.ai (ideally on Pro plan, lite is slow) and BigModel?
(BigModel claims 30-50 t/s)