Google: Gemma 4 31B
google/gemma-4-31b-itAvailable via Safeguard
Private cloudOn-prem
Released
July 1, 2026
Context
256,000 tokens
Input
Price on request
Output
Price on request
Overview
Google open-weight multimodal model (30.7B parameters, text and image in, text out), Apache 2.0 licence.
- Provider
Google
- Modality
- Text and image
- Input
- text, image
- Output
- text
- Weights
- Open weights
- Supported parameters
- tools, reasoning, temperature, top_p, top_kProvider's model page, checked .
- Distillable
- YesLicence or terms, checked .
- Tags
- open-weights, multimodal, vision
Data policy
- Prompts go to
- The host that serves the weights (self-hosted or a hosting provider)
- Retention
- Depends on the host Safeguard routes to. The model maker receives no prompts when the weights are self-hosted
- Zero data retention
- Not stated by the provider
- Region
- Set by the host
Summarised from the provider's published terms. The provider's own policy is what applies.
Where you can run it
How Gemma 4 31B can be deployed, bought and reached through Safeguard.
- Deployment
- Private cloudOn-prem
- Procurement
- Not published yet
- Inference regions
- Not published yet
On-prem deployments run wherever you install them.
Pricing
| Provider | Input / 1M tokens | Output / 1M tokens | Context |
|---|---|---|---|
| Price on request | Price on request | 256K |
- Notes
- Price on request: Google publishes the weights, no first-party per-token price for this size.
- Source
- No per-token list price is published
No markup on model usage: you pay the provider's list price.
API
OpenAI-compatible. Your API key and access to this model are set up for your workspace when you choose Use with Safeguard.
from openai import OpenAI
client = OpenAI(
base_url="https://api.safeguard.sh/gpt/v1",
api_key="SAFEGUARD_API_KEY",
)
response = client.chat.completions.create(
model="google/gemma-4-31b-it",
messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)Base URL https://api.safeguard.sh/gpt/v1, model google/gemma-4-31b-it.
Self-healing security runs on Safeguard.
Your first fix PR is minutes away.
No sales call required, even your agent can complete the purchase over MCP.