News icon

Batch Inference is now generally available on Runpod Serverless

Qwen3.5 4B

Qwen's 4B-parameter open vision-language model with a 256K-token context, ready to serve on Runpod Serverless with vLLM or on a Pod.

Model Information
Provider
Qwen
Modality
Vision + language
Task
Text generation
License
apache-2.0
Released
2026-02-26
Weights on Hugging Face
Qwen/Qwen3.5-4B
Technical Specifications
Parameters
4B
Context
256K tokens
Precision
BF16
Weights
9
Deployment
Recommended starting GPU

RTX 4090

GPU memory

24

Endpoint ID

Qwen3.5 4B

Choose how to deploy your model

Questions? Answers.

The things partners actually ask before they sign, not the things we assume they want to know.

How do I deploy Qwen3.5 4B on Runpod?

Deploy the vLLM worker from the Runpod Hub as a Serverless endpoint and set MODEL_NAME to Qwen/Qwen3.5-4B. The endpoint is OpenAI-compatible and scales to zero when idle. You can also run vLLM on a Pod.

What GPU do I need to run Qwen3.5 4B?

Start with RTX 4090 (24 GB VRAM). The weights take about 9 GB; longer contexts and more concurrent requests need extra memory for the KV cache.

How much does Qwen3.5 4B cost on Runpod?

Serverless endpoints and Pods bill per second for the GPU you choose. A Serverless endpoint costs nothing while scaled to zero.

qwen-qwen3-5-4b
Qwen/Qwen3.5-4B
Vision + language
2026-02-26 16:00
RTX 4090