Tag: Context Management
4 posts
We ran 94,000 trials to measure the cost of harness choice
We analyzed 94,000 benchmark trials to measure how harness choice affects token usage, estimated costs, and accuracy with Claude, GPT, and open-weight models.
Our production SDK hit 99.95% on ARC-AGI-3
Strands Agents scored 99.95% on ARC-AGI-3 using the same minimal harness that developers ship to production every day
Figma to Code at Scale: Building with Strands Agents
A team at Amazon Ads built a multi-agent system that converts Figma designs into production ad layouts. Here's what their architecture looks like, why they chose Strands, and the lessons they learned shipping it.
Reduced cost, better isolation, and more resilience: Strands Agents evolves next-gen capabilities
Strands Agents ships context management that cuts costs in half, Strands Shell for sandboxed agent execution, and chaos testing and red teaming in Strands Evals 1.0.