
Amazon EC2 Inf2 instances, optimized for generative AI inference with up to 40% better price performance, are now generally available in US West (Oregon) region.


<div><p>This article announces the general availability of Amazon EC2 Inf2 instances in the US West (Oregon) region, optimized for generative AI inference workloads.</p><ul><li>Inf2 instances now available in US West (Oregon) region</li><li>Optimized for generative AI applications like text summarization and code generation</li><li>Deliver up to 2.3 petaflops with 384 GB accelerator memory</li><li>Offer 40% better price performance than comparable EC2 instances</li><li>Support scale-out distributed inference via NeuronLink interconnect</li><li>Compatible with popular ML frameworks through AWS Neuron SDK</li><li>Available in four sizes across three AWS regions</li><li>Support On-Demand, Reserved, Spot Instances, and Savings Plans</li></ul><p>Inf2 instances provide cost-effective, high-performance inference for generative AI models with native framework integration and flexible deployment options.</p></div>


Related articles