Multi-model support
Vertesia provides out-of-the-box connectivity for hundreds of generative AI models across all of the leading inference providers, and with our universal prompting, you can write one set of prompts which is instantly usable across any supported model
We believe that multi-model support is foundational to your success with generative AI (GenAI)
In our experience, there is a right model for every generative AI (GenAI) task and activity
For optimal results with complex content, workflows, or agents, you might need to employ multiple models and sequence runs across different models and even different providers.
Why is multi-model so important?
Cost optimization
For some tasks and activities, some models are just more expensive than others. So it is critical that you have a choice in which models you work with and that you can seamlessly switch between models.
Load balancing & failover
Vertesia is the only company to offer virtualized environments that can balance workloads across different models and providers. We also support multi-model failover, enabling you to automatically switch models or providers in the event that your current model is unavailable or is returning erroneous results.
Prevent vendor lock-in
We abstract the underlying GenAI model from your app, service, or agent. This approach not only ensures that you are working with the optimal model but also avoids vendor lock-in. making it painless to adopt new inference models as they become available.
How do you know which GenAI model will deliver the best results?
We provide our own GenAI-driven agent in Vertesia Studio to help you quickly determine the optimal model or models for your use case.
Also, our Services team is here to help guide you along the way
Currently supported inference providers and generative AI models
Inference providers host, and sometimes build their own, generative AI (GenAI) models, including large language models (LLMs) and small language models (SLMs), and are continuously adding new models to their libraries
Vertesia adds new models as soon as they become available, ensuring that our customers get instant access to the latest GenAI model advancements.
Amazon Bedrock
Amazon Bedrock provides their own foundation models (Nova & Titan) plus access to additional models from other providers. Our open-source connector provides access to all available foundation models within Amazon Bedrock.
NOTE:
Below is just a sample of available model providers within the Amazon Bedrock environment in Vertesia Studio. As new models are released, they are made available in Vertesia. If you have questions about our support for a particular model, please contact us.
Google Vertex AI
Google provides their own foundation models (Gemini, Gemma, Imagen, etc.) plus access to additional models from other providers. Our open-source connector to Google Vertex AI provides access to available foundation models within Google Vertex AI.
NOTE:
Below is just a sample of available model providers within the Google Vertex AI environment in Vertesia Studio. As new models are released, they are made available in Vertesia. If you have questions about our support for a particular model, please contact us.
Groq
Groq is a fully integrated inference provider and execution environment in which language models run. Groq created the first Language Processing Unit™ (LPU™) Inference Engine. It is an end-to-end processing unit system that provides extremely fast inference for computationally intensive applications with a sequential component to them.
NOTE:
Below is just a sample of available model providers within the Groq environment in Vertesia Studio. As new models are released, they are made available in Vertesia. If you have questions about our support for a particular model, please contact us.
Hugging Face
We support Hugging Face inference endpoints. Developers can easily deploy transformers, diffusers or any model on dedicated, fully managed infrastructure.
NOTE:
Below is just a sample of available model providers within the Hugging Face environment in Vertesia Studio. As new models are released, they are made available in Vertesia. If you have questions about our support for a particular model, please contact us.
IBM watsonx™
The IBM watsonx™ foundation models library is available in Vertesia and provides access to all available foundational models in the library, including Granite, IBM's flagship series of LLM foundation models based on decoder-only transformer architecture.
NOTE:
Below is just a sample of available model providers within the IBM watsonx™ foundation models library environment in Vertesia Studio. As new models are released, they are made available in Vertesia. If you have questions about our support for a particular model, please contact us.
NOTE:
Below is just a sample of available model providers within the Replicate environment in Vertesia Studio. As new models are released, they are made available in Vertesia. If you have questions about our support for a particular model, please contact us.
NOTE:
Below is just a sample of available model providers within the Together AI environment in Vertesia Studio. As new models are released, they are made available in Vertesia. If you have questions about our support for a particular model, please contact us.
Easily connect to inference providers
In Vertesia, environments are where you connect to inference providers. Simply add your API key to connect to any of the major providers and access their GenAI foundation models using our open-source connectors.
Frequently Asked Questions
Does Vertesia host models?
No, we do not host generative AI models. We instead provide unified access to leading inference providers who host models and to model builders who host their own models, allowing our customers to leverage hundreds of models from different providers to easily select the right model for their particular use case.
Can I use a custom model?
Of course. Vertesia enables customers to connect to their own custom models. We also provide functionality to assist customers in fine-tuning custom models for optimal performance.
What is the difference between a model builder and an inference provider?
Model builders are companies that create and train AI models, including the underlying algorithms and data sets. Inference providers are companies that provide the infrastructure and access to run the AI models. Inference providers often create and train their own models, but inference providers may also operate models from other model providers.
Inference providers that only host models include:
- Groq
- Microsoft (Azure)
- Replicate
- Together AI
Inference providers that host & build models include:
- OpenAI
- Amazon
- IBM
- Mistral AI
Model builders that do not host their models include:
- Meta
- DeepSeek
- Anthropic
- Cohere
Are you intrigued? Let's schedule a demo!
Many organizations don’t understand the full potential of GenAI or how they can transform business processes. Once they see our platform in action, they often have an "aha!" moment, realizing how we can solve real-world challenges.
See how our platform accelerates IT and business outcomes with a live, personalized demo.