Same Model, Different Harness: CUGA, Hermes, and Deep Agents — When to Pick CUGA?
4.32T1.5 sourceHugging Face Blog
Source record
Published by Hugging Face Blog (T1.5 source). The original is at https://huggingface.co/blog/ibm-research/cuga-harness.
Pipeline notes
The summary and note below are generated by the signal pipeline — they are Beyond Desk’s reading, not quotations from the source.
SummaryIBM Research compared CUGA, Hermes, and Deep Agents using the same open-weight model across 90 AppWorld tasks. It reports that changing the harness raised completion from 30% to 81%, then positions CUGA for enterprise workloads requiring enforced rules and deployment within customer-controlled or sovereign environments.
Why it mattersThe controlled model swap isolates harness effects, while deployment and policy criteria help readers match an agent architecture to operational constraints.
Cited by
No citations on record.