# Together AI

> Together AI is a San Francisco-based AI acceleration cloud that lets developers and enterprises train, fine-tune, and run open-source generative AI models on a high-performance GPU platform. Backed by $533M+ from General Catalyst, NVIDIA, Salesforce Ventures and others, it has become one of the most prominent challengers to the closed-model status quo.

- **Founded:** 2022
- **Headquarters:** San Francisco, California, United States
- **Founders:** Vipul Ved Prakash (CEO & Co-founder), Ce Zhang (CTO & Co-founder), Chris Ré (Co-founder), Percy Liang (Co-founder), Tri Dao (Chief Scientist & Co-founder)
- **Team size:** 210+
- **Products:** Together Inference, Together Fine-Tuning, Together GPU Clusters, Together Code Interpreter, Together Custom Models
- **Notable:** Co-released RedPajama, one of the largest open pre-training datasets, Home of FlashAttention, the GPU kernel co-authored by Tri Dao that powers most modern LLMs, Day-one inference partner for Meta Llama 3 and Llama 3.1 launches

## Products & services

- **Together Inference** — Serverless and dedicated endpoints for 200+ open-source models with a custom optimized inference stack.
- **Together Fine-Tuning** — LoRA and full fine-tuning APIs that let teams ship custom models from their own data.
- **Together GPU Clusters** — Reserved NVIDIA H100/H200/GB200 supercomputers for training frontier models.
- **Together Code Interpreter** — Sandboxed code execution environment for agents and AI workflows.
- **Together Custom Models** — End-to-end research-as-a-service to pre-train bespoke foundation models.

## Achievements

- Co-released RedPajama, one of the largest open pre-training datasets
- Home of FlashAttention, the GPU kernel co-authored by Tri Dao that powers most modern LLMs
- Day-one inference partner for Meta Llama 3 and Llama 3.1 launches
- Reached unicorn status within two years of founding
- Reported ~6x revenue growth in 2024

## Latest updates

- **2025-02** — Closed $305M Series B led by General Catalyst and Prosperity7 at a $3.3B valuation.
- **2024-10** — Launched dedicated NVIDIA GB200 clusters for frontier-scale training.
- **2024-07** — Day-one inference partner for Meta Llama 3.1, including the 405B model.

## Links

- Website: https://together.ai
- LinkedIn: https://www.linkedin.com/company/togethercomputer
- Twitter/X: https://twitter.com/togethercompute
- GitHub: https://github.com/togethercomputer
- YouTube: https://www.youtube.com/@togethercompute

---

Profile page: https://yespress.io/together-ai
Published by YesPress — https://yespress.io
Last updated: 2026-05-23