Strands Harness

Strands has released an open-source agent harness that runs locally or on cloud providers with just one line of code, offering 28% lower costs than comparable solutions while maintaining equal or better accuracy across benchmarks. The harness includes built-in optimizations like prompt caching, context management, and token truncation to improve efficiency, and works with models from Anthropic, OpenAI, Google, Amazon Bedrock, and local Ollama instances. It significantly outperforms Claude Code on cost efficiency—costing 77% less on some benchmarks—while remaining competitive on accuracy metrics.