Together Computer is a cloud platform designed for training and serving open-source artificial intelligence models. Founded in 2022, the company provides infrastructure and tools that enable developers and enterprises to build, fine-tune, and deploy large language models and other machine learning systems. Together Computer positions itself as an alternative to proprietary AI offerings, emphasizing transparency, customization, and community-driven development.
The platform supports a range of open-source models, including those based on transformer architectures, and offers services such as managed training clusters, inference endpoints, and fine-tuning APIs. By leveraging high-performance computing resources, including GPUs from vendors like AMD and NVIDIA (not listed, but implied), Together Computer aims to reduce the cost and complexity of running large-scale AI workloads.
History and Founding
Together Computer was established in 2022 by a team of researchers and engineers with backgrounds in machine learning and distributed systems. The founders, including Chen Wu and Freddie Sulit, sought to address the challenges of scaling AI models beyond the capabilities of individual organizations. The company quickly attracted attention for its focus on open-source AI, raising significant funding from venture capital firms.
In 2023, Together Computer launched its flagship platform, which included a suite of tools for training and serving models. The platform gained traction among startups and research institutions looking for flexible, cost-effective alternatives to proprietary cloud services.
Technology and Infrastructure
Together Computer's infrastructure is built on a distributed computing architecture that supports both training and inference. The platform utilizes AWS and other cloud providers to offer scalable GPU clusters, with options for AWS Trainium and other specialized hardware. This allows users to run large-scale training jobs without investing in physical infrastructure.
The company also develops open-source software tools, such as the Together Inference Engine, which optimizes model serving for low latency and high throughput. These tools are designed to work with popular frameworks like PyTorch (not listed, but implied) and TensorFlow (not listed, but implied), and support models from Hugging Face (not listed, but implied) and other repositories.
Products and Services
Together Computer offers a range of products and services tailored to different use cases:
- Managed Training: Users can launch training jobs on dedicated clusters, with support for distributed training and checkpointing.
- Inference Endpoints: Deploy models as APIs with automatic scaling and load balancing.
- Fine-Tuning: Tools for adapting pre-trained models to specific domains or tasks.
- Model Library: A curated collection of open-source models, including LLMs like Llama (not listed, but implied) and Mistral (not listed, but implied).
These services are accessible via a web console, CLI, and REST API, making it easy to integrate with existing workflows.
Open-Source Contributions
Together Computer is a strong advocate for open-source AI. The company has released several models and tools under open licenses, including the Together RedPajama dataset, a large-scale corpus for training language models. This dataset, released in 2023, was created to replicate the quality of proprietary datasets while remaining freely available.
In addition, Together Computer has contributed to the development of generative AI research, collaborating with academic institutions and industry partners. The company's efforts have been recognized in the AI community, with its models appearing on leaderboards and benchmarks.
Market Position and Competition
Together Computer operates in a competitive landscape that includes major cloud providers like AWS, Microsoft Azure, and Google Cloud, as well as specialized AI infrastructure companies such as CoreWeave, Cerebras, and Groq. Unlike these competitors, Together Computer focuses exclusively on open-source models, offering a more flexible and customizable platform.
The company also competes with AI research organizations like OpenAI, Anthropic, and Google DeepMind, but differentiates itself by not developing proprietary models. Instead, it empowers users to build and deploy their own models, fostering innovation and reducing dependency on closed systems.
Future Directions
As of 2025, Together Computer continues to expand its offerings, with plans to support more hardware accelerators and improve its software stack. The company is also exploring partnerships with semiconductor companies like AMD and Intel to optimize performance on diverse architectures.
With the growing demand for AI solutions, Together Computer aims to become a leading platform for open-source AI development, enabling a broader community to participate in the advancement of machine learning.