AWS adds Advanced Prompt Optimization tool to Bedrock

"The tool, Amazon Bedrock Advanced Prompt Optimization, can be accessed through the Bedrock console, and is designed to automatically refine prompts for better accuracy, consistency, and efficiency across multiple large language models, the hyperscaler wrote in a blog post."

"The tool works by first evaluating prompts against user-defined datasets and metrics, then rewriting them to optimize them for up to five inference models. It then benchmarks the optimized versions against the originals across the models to help developers identify the best-performing configurations for specific workloads, AWS said."

"The company said that enterprise customers will be billed for its use based on the Bedrock model inference tokens consumed during the optimization process, using the same per-token pricing rates applied to standard Bedrock inference workloads."

""Enterprise demand for such tools is being driven by a convergence of cost pressure [and] operational complexity when it comes to scaling AI, rather than any single factor," said Gaurav Dewan, research director at Avasant."

Amazon Bedrock added Amazon Bedrock Advanced Prompt Optimization to help refine prompts for generative AI applications. The tool is available through the Bedrock console and automatically improves prompts for better accuracy, consistency, and efficiency across multiple large language models. It evaluates prompts using user-defined datasets and metrics, rewrites prompts to optimize them for up to five inference models, and benchmarks optimized versions against the originals across those models. The benchmarking helps developers identify best-performing configurations for specific workloads. The tool is generally available across many AWS regions. Enterprise customers are billed based on Bedrock model inference tokens consumed during optimization, using the same per-token pricing as standard Bedrock inference workloads. Analysts link demand to cost pressure and operational complexity when scaling AI in production.

Read at InfoWorld