• Up to 10k+ Frontier GPU Clusters: Optimized for foundation model training.
• NVIDIA H100 and H200: Massive fleet of Hopper GPUs available now.
• ‍NVIDIA GB200 NVL72: 36 Grace CPUs connected to 72 Blackwell GPUs.
• Together Kernel Collection: Accelerating fundamental AI operations.
• End-to-End Platform: From pre-training, fine-tuning to inference.‍
• Flexible Commits: Starting from one month with scheduled buildup options.‍
• Premium Support: Included with every cluster.‍
• Industry-Leading Research: 9x faster training with FlashAttention-3.

‍

“Together GPU Clusters provided a combination of amazing training performance, expert support, and the ability to scale to meet our rapid growth to help us serve our growing community of AI creators.” — Demi Guo, CEO, Pika Labs