GLM 5 Fp4 is NVFP4 quantized version of Z.AI's 744B MoE model with 40B active params. Runs inference on vLLM or SGLang. Designed for coding and agents.

How does GLM 5 Fp4 API work?

Call GLM 5 Fp4 API endpoint with text prompts up to 200K tokens. Supports tool calling and chunked prefill. Use tensor-parallel-size 8 for multi-GPU.

Is GLM 5 Fp4 model open source?

Base GLM-5 uses MIT license for commercial use. GLM 5 Fp4 from NVIDIA on Hugging Face. Quantized weights public.

GLM 5 Fp4 vs full GLM-5?

GLM 5 Fp4 quantizes to NVFP4, cuts memory vs BF16. Retains 86%+ benchmark scores. Ideal for efficient deployment.

Best GLM 5 Fp4 alternative?

GLM 5 Fp4 leads open weights on agentic index at 63 score. Alternatives like DeepSeek V3 lag on coding benches.

GLM 5 Fp4 LLM context length?

Supports 200K tokens input. Max output around 131K. Handles long-horizon tasks like full codebases.

Available now on ModelsLab · Language Model

GLM 5 Fp4
Quantized Power. Full Scale

Try GLM 5 Fp4 API Documentation

Run GLM 5 Fp4 Efficiently

NVFP4 Quantized

744B MoE Optimized

Activates 40B parameters per token in GLM 5 Fp4 for low-cost inference.

200K Context

Handles Long Tasks

Processes massive codebases and documents with GLM 5 Fp4 model.

Agentic Coding

Native Tool Calling

Supports function execution and planning via GLM 5 Fp4 API.

Examples

See what GLM 5 Fp4 can create

Copy any prompt below and try it yourself in the playground.

Code Refactor

“Refactor this Python function for efficiency, add type hints, and optimize for async execution. Original code: def fetch_data(url): response = requests.get(url); return response.json()”

Agent Plan

“Plan steps to deploy a web app: select stack, write Dockerfile, set CI/CD pipeline, handle scaling with Kubernetes.”

SQL Query

“Write SQL query joining users and orders tables, filter by date range 2025-01-01 to 2026-04-01, group by user_id, sum revenue.”

Debug Script

“Debug this bash script failing on loop: for i in {1..10}; do echo $i >> log.txt; done. Fix permissions and error handling.”

For Developers

A few lines of code.
GLM 5 Fp4. One API call.

ModelsLab handles the infrastructure: fast inference, auto-scaling, and a developer-friendly API. No GPU management needed.

Serverless: scales to zero, scales to millions
Pay per token, no minimums
Python and JavaScript SDKs, plus REST API

API Documentation

import requests

response = requests.post(
    "https://modelslab.com/api/v7/llm/chat/completions",
    json={
  "key": "YOUR_API_KEY",
  "prompt": "",
  "model_id": ""
}
)
print(response.json())

FAQ

Common questions about GLM 5 Fp4

Read the docs

Ready to create?

Start generating with GLM 5 Fp4 on ModelsLab.