Use Gemma open models

Gemma is a set of lightweight, generative artificial intelligence (AI) open models. Gemma models are available to run in your applications and on your hardware, mobile devices, or hosted services. You can also customize these models using tuning techniques so that they excel at performing tasks that matter to you and your users. Gemma models are based on Gemini models and are intended for the AI development community to extend and take further.

Fine-tuning can help improve a model's performance in specific tasks. Because models in the Gemma model family are open weight, you can tune any of them using the AI framework of your choice and the Vertex AI SDK. You can open a notebook example to fine-tune the Gemma model using a link available on the Gemma model card in Model Garden.

The following Gemma models are available to use with Vertex AI. To learn more about and test the Gemma models, see their Model Garden model cards.

Model name	Use cases	Model Garden model card
Gemma 3n	Capable of multimodal input, handling text, image, video, and audio input, and generating text outputs.	Go to the Gemma 3n model card
Gemma 3	Best for text generation and image understanding tasks, including question answering, summarization, and reasoning.	Go to the Gemma 3 model card
Gemma 2	Best for text generation, summarization, and extraction.	Go to the Gemma 2 model card
Gemma	Best for text generation, summarization, and extraction.	Go to the Gemma model card
CodeGemma	Best for code generation and completion.	Go to the CodeGemma model card
PaliGemma 2	Best for image captioning tasks and visual question and answering tasks.	Go to the PaliGemma 2 model card
PaliGemma	Best for image captioning tasks and visual question and answering tasks.	Go to the PaliGemma model card
ShieldGemma 2	Checks the safety of synthetic and natural images to help you build robust datasets and models.	Go to the ShieldGemma 2 model card
TxGemma	Best for therapeutic prediction tasks, including classification, regression, or generation, and reasoning tasks.	Go to the TxGemma model card
MedGemma	Gemma 3 variants that are trained for performance on medical text and image comprehension.	Go to the MedGemma model card
MedSigLIP	SigLIP variant that is trained to encode medical images and text into a common embedding space.	Go to the MedSigLIP model card
T5Gemma	Well-suited for a variety of generative tasks, including question answering, summarization, and reasoning.	Go to the T5Gemma model card

The following are some options for where you can use Gemma:

Use Gemma with Vertex AI

Vertex AI offers a managed platform for rapidly building and scaling machine learning projects without needing in-house MLOps expertise. You can use Vertex AI as the downstream application that serves the Gemma models. For example, you might port weights from the Keras implementation of Gemma. Next, you can use Vertex AI to serve that version of Gemma to get predictions. We recommend using Vertex AI if you want end-to-end MLOps capabilities, value-added ML features, and a serverless experience for streamlined development.

To get started with Gemma, see the following notebooks:

• Serve Gemma 3n in Vertex AI

• Serve Gemma 3 in Vertex AI

• Serve Gemma 2 in Vertex AI

• Serve Gemma in Vertex AI

• Fine-tune Gemma 3 using PEFT and then deploy to Vertex AI from Vertex

• Fine-tune Gemma 2 using PEFT and then deploy to Vertex AI from Vertex

• Fine-tune Gemma using PEFT and then deploy to Vertex AI from Vertex

• Fine-tune Gemma using PEFT and then deploy to Vertex AI from Huggingface

• Fine-tune Gemma using KerasNLP and then deploy to Vertex AI

• Fine-tune Gemma with Ray on Vertex AI and then deploy to Vertex AI

• Run local inference with ShieldGemma 2 with Hugging Face transformers

• Serve MedGemma in Vertex AI

• Serve MedSigLIP in Vertex AI

• Run local inference with T5Gemma with Hugging Face transformers

Use Gemma in other Google Cloud products

You can use Gemma with other Google Cloud products, such as Google Kubernetes Engine and Dataflow.

Use Gemma with GKE

Google Kubernetes Engine (GKE) is the Google Cloud solution for managed Kubernetes that provides scalability, security, resilience, and cost effectiveness. We recommend this option if you have existing Kubernetes investments, your organization has in-house MLOps expertise, or if you need granular control over complex AI/ML workloads with unique security, data pipeline, and resource management requirements. To learn more, see the following tutorials in the GKE documentation:

• Serve Gemma with vLLM
• Serve Gemma with TGI
• Serve Gemma with Triton and TensorRT-LLM
• Serve Gemma with JetStream

Use Gemma with Dataflow

You can use Gemma models with Dataflow for sentiment analysis. Use Dataflow to run inference pipelines that use the Gemma models. To learn more, see Run inference pipelines with Gemma open models.

Use Gemma with Colab

You can use Gemma with Colaboratory to create your Gemma solution. In Colab, you can use Gemma with framework options such as PyTorch and JAX. To learn more, see:

• Get started with Gemma using Keras.
• Get started with Gemma using PyTorch.
• Basic tuning with Gemma using Keras.
• Distributed tuning with Gemma using Keras.

Gemma model sizes and capabilities

Gemma models are available in several sizes so you can build generative AI solutions based on your available computing resources, the capabilities you need, and where you want to run them. Each model is available in a tuned and an untuned version:

• Pretrained - This version of the model wasn't trained on any specific tasks or instructions beyond the Gemma core data training set. We don't recommend using this model without performing some tuning.

• Instruction-tuned - This version of the model was trained with human language interactions so that it can participate in a conversation, similar to a basic chat bot.

• Mix fine-tuned - This version of the model is fine-tuned on a mixture of academic datasets and accepts natural language prompts.

Lower parameter sizes means lower resource requirements and more deployment flexibility.

Model name	Parameters size	Input	Output	Tuned versions	Intended platforms
Gemma 3n
Gemma 3n E4B	4 billion effective parameters	Text, image and audio	Text	Pretrained Instruction-tuned	Mobile devices and laptops
Gemma 3n E2B	2 billion effective parameters	Text, image and audio	Text	Pretrained Instruction-tuned	Mobile devices and laptops
Gemma 3
Gemma 27B	27 billion	Text and image	Text	Pretrained Instruction-tuned	Large servers or server clusters
Gemma 12B	12 billion	Text and image	Text	Pretrained Instruction-tuned	Higher-end desktop computers and servers
Gemma 4B	4 billion	Text and image	Text	Pretrained Instruction-tuned	Desktop computers and small servers
Gemma 1B	1 billion	Text	Text	Pretrained Instruction-tuned	Mobile devices and laptops
Gemma 2
Gemma 27B	27 billion	Text	Text	Pretrained Instruction-tuned	Large servers or server clusters
Gemma 9B	9 billion	Text	Text	Pretrained Instruction-tuned	Higher-end desktop computers and servers
Gemma 2B	2 billion	Text	Text	Pretrained Instruction-tuned	Mobile devices and laptops
Gemma
Gemma 7B	7 billion	Text	Text	Pretrained Instruction-tuned	Desktop computers and small servers
Gemma 2B	2.2 billion	Text	Text	Pretrained Instruction-tuned	Mobile devices and laptops
CodeGemma
CodeGemma 7B	7 billion	Text	Text	Pretrained Instruction-tuned	Desktop computers and small servers
CodeGemma 2B	2 billion	Text	Text	Pretrained	Desktop computers and small servers
PaliGemma 2
PaliGemma 28B	28 billion	Text and image	Text	Pretrained Mix fine-tuned	Large servers or server clusters
PaliGemma 10B	10 billion	Text and image	Text	Pretrained Mix fine-tuned	Higher-end desktop computers and servers
PaliGemma 3B	3 billion	Text and image	Text	Pretrained Mix fine-tuned	Desktop computers and small servers
PaliGemma
PaliGemma 3B	3 billion	Text and image	Text	Pretrained Mix fine-tuned	Desktop computers and small servers
ShieldGemma 2
ShieldGemma 2	4 billion	Text and image	Text	Fine-tuned	Desktop computers and small servers
TxGemma
TxGemma 27B	27 billion	Text	Text	Pretrained Instruction-tuned	Large servers or server clusters
TxGemma 9B	9 billion	Text	Text	Pretrained Instruction-tuned	Higher-end desktop computers and servers
TxGemma 2B	2 billion	Text	Text	Pretrained	Mobile devices and laptops
MedGemma
MedGemma 27B	27 billion	Text and image	Text	Text-only instruction-tuned Instruction-tuned	Large servers or server clusters
MedGemma 4B	4 billion	Text and image	Text	Pretrained Instruction-tuned	Desktop computers and small servers
MedSigLIP
MedSigLIP	800 million	Text and image	Embedding	Fine-tuned	Mobile devices and laptops
T5Gemma
T5Gemma 9B-9B	18 billion	Text	Text	PrefixLM, pretrained PrefixLM, instruction-tuned UL2, pretrained UL2, instruction-tuned	Mobile devices and laptops
T5Gemma 9B-2B	11 billion	Text	Text	PrefixLM, pretrained PrefixLM, instruction-tuned UL2, pretrained UL2, instruction-tuned	Mobile devices and laptops
T5Gemma 2B-2B	4 billion	Text	Text	PrefixLM, pretrained PrefixLM, instruction-tuned UL2, pretrained UL2, instruction-tuned	Mobile devices and laptops
T5Gemma XL-XL	4 billion	Text	Text	PrefixLM, pretrained PrefixLM, instruction-tuned UL2, pretrained UL2, instruction-tuned	Mobile devices and laptops
T5Gemma M-L	2 billion	Text	Text	PrefixLM, pretrained PrefixLM, instruction-tuned UL2, pretrained UL2, instruction-tuned	Mobile devices and laptops
T5Gemma L-L	1 billion	Text	Text	PrefixLM, pretrained PrefixLM, instruction-tuned UL2, pretrained UL2, instruction-tuned	Mobile devices and laptops
T5Gemma B-B	0.6 billion	Text	Text	PrefixLM, pretrained PrefixLM, instruction-tuned UL2, pretrained UL2, instruction-tuned	Mobile devices and laptops
T5Gemma S-S	0.3 billion	Text	Text	PrefixLM, pretrained PrefixLM, instruction-tuned UL2, pretrained UL2, instruction-tuned	Mobile devices and laptops

Gemma has been tested using Google's purpose built v5e TPU hardware and NVIDIA's L4(G2 Standard), A100(A2 Standard), H100(A3 High) GPU hardware.

What's next

• See Gemma documentation.