Amazon Bedrock Advanced Prompt Optimization now supports prompt shortening

Amazon Bedrock Advanced Prompt Optimization now supports prompt shortening, an optional setting that reduces prompt length while maintaining or improving evaluation scores.

• Reduces prompt length to lower input token costs and decrease time to first token
• Returns optimized prompt only if shorter than original and scores at least as high
• Provides compression ratio and original/optimized token counts in results
• Works with any evaluation method: steering criteria, LLM-as-a-judge rubrics, custom Lambda evaluators
• Available in all AWS Regions where Advanced Prompt Optimization is available

Prompt shortening helps optimize costs and latency for large-scale and latency-sensitive applications without sacrificing quality.