LongCat-2.0 Free

LongCat-2.0 Free sits within the longcat family and is positioned in third-party marketplaces as a live, freely accessible model intended for high-throughput coding and agent workflows. Independent listing evidence confirms that it appears alongside GLM 5.3 and DeepSeek V4 in an open-source-plus-third-party inference catalog, suggesting it is meant for general text-based reasoning tasks where a large working window is helpful for codebase-scale context.

The model is described by the same third-party host as operating with a one-the cataloged API limit, an unusually large working memory that fits long-document and whole-codebase agent scenarios rather than short conversational turns. Because creator-level documentation such as a model card, technical report, or benchmark disclosure was not part of the supplied evidence, the qualitative picture is limited to its marketplace placement and context scale, leaving architecture lineage, training scale, and benchmark standing open to future confirmation.