LMIS-ORG/MemAgent_Slime_Agentic_Qwen2.5_7B
TEXT GENERATIONConcurrency Cost:1Model Size:7.6BQuant:FP8Ctx Length:32kPublished:Apr 2, 2026Architecture:Transformer0.0K Cold

LMIS-ORG/MemAgent_Slime_Agentic_Qwen2.5_7B is a 7.6 billion parameter language model developed by LMIS-ORG, based on the Qwen2.5 architecture. This model is specifically designed for processing arbitrarily long documents by compressing them into a fixed-size recurrent memory through a chunk-by-chunk LLM update loop. It excels at question answering over extended contexts, demonstrating superior performance on the RULER-HQA benchmark across context lengths up to 448K tokens compared to larger baseline models.

Loading preview...