Gemma-4-31B¶
In this article
Information
Gemma 4-31B is the largest and most powerful model in the Gemma 4 family, designed to solve the most complex tasks. Unlike the MoE version, this model uses a classic Dense architecture with 30.7 billion parameters, providing exceptional stability and reasoning depth. Deployment is based on Ubuntu 22.04 and requires significant computational resources: full operation requires approximately 62 GB of VRAM; however, with 4-bit quantization, the model runs successfully on systems with 20 GB of memory. Integration with Open WebUI provides a convenient interface for interacting with the model while maintaining full control over data.
Key Features¶
- Powerful Dense Architecture: The model is a leader in the dense open-weights category, offering the highest quality generation and analysis;
- Native Multimodality: The architecture supports simultaneous processing of text and image input with text output;
- Massive Context Window: Support for up to 256K tokens allows loading entire code repositories or long documents into memory;
- Thinking Mode: The model is capable of generating intermediate reasoning steps before formulating the final answer;
- Agentic Capabilities: Built-in support for native function calling simplifies the creation of autonomous AI agents;
- Open WebUI Integration: Provides a modern web interface for convenient interaction with the model via port 8080, ensuring full control over data and request processing;
- Scalability: Support for multi-GPU configurations and the ability to distribute load across several GPUs for optimal performance;
- Security and Control: Full local deployment ensures data privacy, while
OLLAMA_HOSTandOLLAMA_ORIGINSsettings guarantee network security; - Performance: Uses
LLAMA_FLASH_ATTENTIONtechnology to accelerate request processing and optimize model operation; - Fault Tolerance: A built-in automatic restart system for containers and services ensures stable operation.
Deployment Features¶
| ID | Software Name | OS | VPS | BM | VGPU | GPU | Min. CPU (Cores) | Min. RAM (GB) | Min. HDD/SSD (GB) | Custom Domain | Available |
|---|---|---|---|---|---|---|---|---|---|---|---|
Information
Deployment details for this software can be found at this link
Note
Unless otherwise specified, we install the latest release version of the software from the developer's website or operating system repositories by default.
Getting Started After Deploying Gemma-4-31B¶
After payment is completed, a notification regarding server readiness will be sent to the email address provided during registration. It will contain the VPS IP address, login and password for connecting to the server, and a link to access the Open WebUI panel. Customers manage their hardware through the Server Control Panel and API — Invapi.
-
Authorization credentials for accessing the server operating system (e.g., via SSH) will be sent to you in an email.
-
Link to access the Ollama control panel with Open WebUI web interface: located under the webpanel tag in the Configuration >> Tags tab of the Invapi control panel. The exact link, formatted as
https://gemma31b<Server_ID_from_Invapi>.hostkey.com, is sent in the email provided upon server delivery.
After clicking the link from the webpanel tag, the Get started with Open WebUI authorization window will open, where you must create an administrator username, email, and password for the chatbot, and then click the Create Admin Account button:
Attention
After the first user registers, the system automatically assigns them the Administrator role. To ensure security and control over the registration process, all subsequent registration requests must be approved in the Open WebUI interface from an administrator account.
Note
A detailed description of how to work with the Ollama control panel via Open WebUI can be found in the AI Chatbot on Your Own Server article.
Note
For optimal performance, it is recommended to use a GPU with more than the minimum requirement of 24 GB VRAM, which provides headroom for processing large contexts and parallel requests. Detailed information regarding core Ollama and Open WebUI settings can be found in the Ollama developer documentation and the Open WebUI developer documentation.
Ordering a Server with Gemma-4-31B via API¶
To install this software using the API, follow these instructions.