GLM-5.3-Flash

Z-AI GLM-5.3 Flash 320B MoE (18B active) on NVIDIA NIM — hybrid KDA attention, native FP8, 1M context, 40 RPM.

Updated 11/22/20261 code examples

About the Model

为什么选择 GLM 5.3 Flash?

GLM 5.3 Flash 是智谱 AI 推出的高速中英双语基座大模型,具备极佳的长文总结与复杂逻辑分析能力。

👉 注册福利:通过专属推荐链接 https://www.vultr.com/?ref=9923295-9J 注册并绑定有效支付方式过验,即可立享 $300 免费试用额度(30天有效),直接抵扣 Serverless Inference 推理费用。

How to Access for Free (via NVIDIA NIM)

Why Choose GLM-5.3-Flash on NVIDIA NIM?

GLM-5.3-Flash from Zhipu AI is hosted on NVIDIA NIM with native FP8 weights and TensorRT-LLM acceleration. Featuring a 320B MoE architecture with 18B active parameters, hybrid KDA attention, and 1M context, it delivers industry-leading Chinese reasoning, code synthesis, and multimodal comprehension.

Try it in your browser

Pick a model, paste your own free key, and run. Your key is sent once to call the provider and never stored on our servers.

Pick a model, paste your own free key, and run. Your key is sent once to call the provider and never stored on our servers.

Code Examples

curl
curl https://integrate.api.nvidia.com/v1/chat/completions \
  -H "Authorization: Bearer $NVIDIA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "z-ai/glm-5.3-flash",
    "messages": [{"role": "user", "content": "请分析分布式系统脑裂问题并给出基于 Raft 的解决方案。"}]
  }'

Community Feedback

Comments are tied to your GitHub account — sign in to join the discussion.