Together AI

Together AI supports a wide range of state-of-the-art open-source models including Llama 3, Mixtral, Qwen, and Stable Diffusion. We constantly update our library to include the latest releases, allowing you to run inference on high-performance models without managing your own infrastructure.

Yes, Together AI provides a dedicated fine-tuning API that allows you to customize open-source models with your proprietary data. Our infrastructure is optimized for efficient training, enabling you to create specialized versions of models like Llama for your specific business use cases.

While the Inference API provides serverless access to models, GPU Clusters offer dedicated H100 or A100 instances for large-scale training and research. These clusters are interconnected with high-speed networking, making them ideal for organizations building their own foundational models from scratch.

You can get started in minutes by creating an account, generating an API key, and using our playground to test prompts. Our serverless infrastructure handles the deployment automatically, so there is no need to manually provision servers or manage complex Kubernetes clusters.

We provide comprehensive documentation, migration guides, and code samples to help you move your workloads from local hardware to our cloud. Our support team can also assist in optimizing your inference parameters to ensure you achieve the best performance-to-cost ratio.

Together AI offers an OpenAI-compatible API, meaning you can often switch your existing integrations by simply changing the base URL and API key. This allows developers to migrate their applications to open-source models with minimal code changes and no architectural overhaul.

We provide official SDKs for Python and TypeScript/JavaScript to streamline your development process. Additionally, because we use a standard RESTful API, you can integrate Together AI services into any environment that supports HTTP requests, including Go, Ruby, and Java.

Our Inference API uses a transparent pay-as-you-go model based on the number of tokens processed (for LLMs) or images generated. This allows you to scale your application without upfront costs, only paying for the exact volume of requests your users generate each month.

Yes, we offer custom enterprise pricing and committed use discounts for customers with high-volume requirements. If your application processes millions of tokens daily or requires dedicated capacity, our sales team can structure a plan that reduces your effective cost per request.

We prioritize data privacy and do not use your input data or output generations to train our base models. All data is encrypted in transit using TLS, and we offer enterprise-grade security features to ensure that your proprietary information remains confidential during processing.

Together AI is committed to high security standards and maintains SOC 2 Type II compliance. We provide the necessary documentation and security controls required by enterprise legal and IT departments to ensure our infrastructure meets rigorous data protection requirements.

All users have access to our extensive documentation, community forums, and email support. Enterprise customers receive enhanced support packages, which include dedicated Slack channels, faster response times, and direct access to our engineering team for architectural reviews.