Google: the Gemma family of open models

Run Google's lightweight open models, built from the same research as Gemini.

Benefits of running Google on Crusoe Cloud

Serve more users on the same GPUs

Crusoe’s inference engine is powered by MemoryAlloy™ technology, a cluster-wide KV cache that pairs with cache-aware routing to send each request to the node most likely to have its context ready. Gemma models handle more concurrent traffic at consistent low latency, without adding capacity to keep up.

Make Gemma your own

Run a LoRA fine-tuning job on your data with Serverless Fine-Tuning, deploy the result in one click, or download your weights in .safetensors format. With Gemma’s Apache 2.0 license, what you build is yours to run anywhere.

Go from first call to production with ease

Gemma is one of the few model families on Crusoe available across all three services: Serverless Inference, Self-Serve Deployments, and Serverless Fine-Tuning. Prototype, tune, and scale to a dedicated deployment without switching models or platforms.

All Google models

Text
Vision
Self-serve
Google

Gemma 4 26B A4B

Context
262144
Text
Vision
Featured
Self-serve
Fine-tuning
Serverless
Google

Gemma 4 31B

Context
262144

Fine-tune and deploy in Crusoe Intelligence Foundry

Serverless
Inference

Get started fast with usage-based consumption of top open models via a fully managed inference API.

Best for
• Early-stage workloads 
• Low volume traffic
• Rapid experimentation

Self-Serve Deployments

Serve models with a dedicated endpoint, optimized for throughput or responsiveness.

Best for
• Production-scale workloads
• Low-latency applications
• Predictable costs

Tailored
Deployments

Work with our team for the highest level of optimization and a dedicated, benchmarked endpoint. You bring the model and define the requirements, we handle the rest.

Best for
• Proprietary models
• Bespoke inference optimization
• SLA-backed performance

Serverless
Fine-Tuning

Quickly adapt models to your domain. You bring the dataset and pick a base model. We handle the GPUs, the training job and the checkpoints. Go live with Self-Serve Deployments in one click, or download your weights to deploy anywhere.

Best for
• Domain-specific accuracy
• Consistent output formatting
• Smaller models, lower cost

Frequently
asked questions

Are you ready to build something amazing?