Languagetext

GLM 5

GLM-5 is Zhipu AI's (Z.AI) flagship foundation model launched February 11, 2026, a 744B-parameter mixture-of-experts model with about 40B active parameters, roughly 200K context, pre-trained on 28.5T tokens using DeepSeek Sparse Attention. It centers on agentic intelligence, autonomously decomposing objectives into subtasks, with reported 77.8% on SWE-bench Verified, and was trained on domestically produced Huawei Ascend chips. Released as open weights under the MIT License. Ideal for agentic coding and multi-step tool-chaining workloads where open-weight self-hosting is desired.

Input

Prompt

Top p

Thinking

Temperature

Max completion tokens

1200000

Ready to generate

API Reference

POSThttps://api.token360.ai/v1/chat/completions

curl https://api.token360.ai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer your-api-key" \
  -d '{
  "model": "glm-5",
  "messages": [
    {
      "role": "user",
      "content": "Hello!"
    }
  ]
}'