Vision
How to deploy gemma-4-31B-it on a GPU cloud
A 31.3B vision-language model that reads images alongside text. Full specs, license and use cases.
gemma-4-31B-it size and hardware requirements
31.3B
Total parameters
Dense (no MoE)
Architecture
BF16
Published precision
69.9 GB
Min VRAM (native)
| Precision | Weight size on disk | Required VRAM | Cheapest live fit | GPUs needed | Est. $/hr (full fit) |
|---|---|---|---|---|---|
| BF16 | 58.3 GB | 69.9 GB | A100 | 1 | $0.851/hr |
| FP8 (quantized) | 29.1 GB | 35.0 GB | RTX 6000 Ada | 1 | $0.524/hr |
| INT4 (quantized) | 14.6 GB | 17.5 GB | RTX 3090 | 1 | $0.147/hr |
How to run gemma-4-31B-it
Run gemma-4-31B-it with vLLM
Generic example, not from the model's own docs: adjust flags (quantization, context length, parallelism) for your setup.
vllm serve google/gemma-4-31B-it --tensor-parallel-size 1Run gemma-4-31B-it with Ollama
Verified against Ollama's own library listing.
ollama run gemma4:31bRun gemma-4-31B-it with GGUF quantizations
Prebuilt GGUF weights published at unsloth/gemma-4-31B-it-GGUF. Run with llama.cpp's llama-server or load the repo directly in LM Studio.
llama-server -hf unsloth/gemma-4-31B-it-GGUFDeploy gemma-4-31B-it on Aquanode
Aquanode has no one-click deploy template for gemma-4-31B-it; you install the inference engine yourself with the commands below. Aquanode sells GPU pods billed per second, not a hosted inference API.
- Launch a bare GPU pod sized to the requirement above (1× A100 or larger).
- Open a terminal on the pod, or save one of the commands above as a startup script so it runs automatically the first time the pod boots.
- Run the command and connect to the resulting endpoint.
Launch a GPU pod