Confidential LLM Inference: Performance and Cost Across CPU and GPU TEEs

Open in new window