Strands harness is a new open-source agent framework that runs locally or in the cloud with 28% lower token costs than competing solutions while maintaining comparable accuracy. Built on the Strands Harness SDK with integrated prompt caching and context management, it supports multiple model providers and comes with shell, file, and web tools out of the box.
A comparative study tested three context compaction strategies—FutureOS, OpenCode, and Codex—on their ability to retain information from agent sessions. FutureOS retained 83% of queryable information, significantly outperforming OpenCode (47%) and Codex (38%), with the key difference being that FutureOS preserves assistant prose while others compress it away. The analysis reveals that tool output dominates context volume but is rarely referenced, while the sparse assistant text is the primary source of follow-up questions.
Strands harness is a new open-source agent framework that achieves 28% lower token costs than competing solutions while maintaining equal or better accuracy across benchmarks. Available for Python and TypeScript, it runs locally or on cloud providers with built-in prompt caching, context management, and support for multiple model providers including Claude, GPT, and Deepseek.
Researchers at NVIDIA, NTU, and MIT developed SoL-Pi, a system that optimizes coding agents through recursive auto-research loops across diverse environments. SoL-Pi reduces token usage by 44.7-49.0% and API costs by approximately one third while maintaining performance comparable to existing systems on the 51-task EdgeBench evaluation.
Grok 4.7 is launching soon with 2.1 trillion parameters as a new pretraining effort rather than a refresh of 4.6. The model outperforms 4.6 in most areas with improved token efficiency and benefits from SpaceX and Starlink engineering data, though multimodal capabilities need further development and serving speed is slightly slower.