Understand this story
On October 21, 2026, Databricks released MemCurrent, a persistent memory architecture for production agents. The architecture combines episodic memory with derived memory and offers three integration methods: Plugin, Tool, and offline dreaming. According to the same report, based on 114 internal test cases selected from more than 12,000 real Genie sessions, the pass rate rose from 22.3% without memory to 68.4%, while being 1.35 times faster and 1.44 times cheaper; on the public benchmarks LoCoMo, LongMemEval, and ConvoMem, it averaged a score of 88.1%, comparable to 87.8% for full context, but with about 12 times fewer input tokens per question (8.4K versus 103K). The report also states that the architecture has been used in Agent Bricks Memory and Genie. There are currently no subsequent independent development facts.
AI-generated from public reports
Latest development · Report
Databricks Releases MemCurrent Long-Term Memory Architecture, Internal Benchmark Pass Rate TripledDevelopments
1 confirmed developments
- Source:Shao Meng
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Oct 21
Shao Meng
Databricks Releases MemCurrent Long-Term Memory Architecture, Internal Benchmark Pass Rate Tripled
Databricks has released MemCurrent, a persistent memory architecture for production Agents, combining episodic memory with derived memory and adopting three integration approaches: Plugin, Tool, and offline dreaming. Based on 114 internal test cases selected from more than 12,000 real Genie sessions, the pass rate rose from 22.3% without memory to 68.4%, while also being 1.35 times faster and 1.44 times cheaper. On the public benchmarks LoCoMo, LongMemEval, and ConvoMem, the average score was 88.1%, comparable to 87.8% with full context, but with about 12 times fewer input tokens per question (8.4K vs. 103K). The architecture has already been used in Agent Bricks Memory and Genie.