Harness Engineering, Not Models, Is the Agent Bottleneck: Study

Rajiv Shah argued at ODSC that harness engineering, not model capability, drives agent performance: swapping harnesses lifted 7 models' average score from 23.3% to 62.4% while cutting costs by 60%.

2026-10-06 ~ 2026-10-06 · 2 related posts

• Same models, new harness: scores jump 23% to 62%, says Rajiv Shah — rajistics · 2026-10-06
• Harvey's rebuilt harness lifts 7-model average from 23.3% to 62.4% while cutting cost 60% — rajistics · 2026-10-06