Serverless GPUs Public Preview: Run AI workloads on H100, A100, L40S, and more
Blog post from Koyeb
Koyeb has announced the public preview of GPU-accelerated workloads, introducing high-performance GPUs like the H100 and A100, which are ideal for generative AI processes such as large language models and image generation. The GPUs, ranging from 20GB to 80GB of vRAM, offer support for high-precision calculations and varying bandwidth capabilities, with on-demand pricing from $0.50/hr to $3.30/hr. Koyeb's platform provides a seamless serverless deployment experience, including features like one-click container deployment, built-in load balancing, horizontal autoscaling, zero downtime deployments, and real-time monitoring. Users can get started easily through the Koyeb control panel or CLI, and the platform offers $200 of credit for new users. Future plans include enhancing autoscaling performance and expanding GPU offerings to meet diverse user needs.