Five AI agent harnesses took the same memory test. The results weren’t even close.
AI agents are getting better at remembering past conversations and earlier tasks but remembering something doesn’t guarantee an agent can The post Five AI agent harnesses took the same memory test. The results weren’t even close. appeared first on The New Stack .
Five AI agent harnesses competed in a memory test, with the results showing a significant gap between them. Letta Dream scored the highest at 73.3%, followed by Prime Agent at 60%. OpenAI Agents SDK lagged behind at 40%, while Claude Managed Agents and Hermes Agent trailed at 33.3% and 26.7%, respectively. According to Letta co-founders Charles Packer and Sarah Wooders, the gap highlights a gap between simply recalling past information and using it to improve future work.
Letta's FrontierMemory benchmark measures how well agents can learn from experience, rather than just retrieving information stored from previous tasks. The benchmark was designed to test whether agents can learn and improve their performance over time. The results suggest that the harness, or the infrastructure managing how an agent operates, may play a crucial role in an agent's ability to learn and apply lessons from previous tasks.
Written by urgent.news from The New Stack's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
