specialv/Vims2-7B
TEXT GENERATIONConcurrency Cost:1Model Size:7.6BQuant:FP8Ctx Length:32kPublished:Mar 29, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

specialv/Vims2-7B is a 7.6 billion parameter large language model based on the Qwen 2.5 architecture, developed by specialv using the Task Arithmetic merging method. It specializes in logical reasoning, mathematical problem-solving, and coding, while maintaining strong instruction-following capabilities in both Italian and English. With a context window of 131,072 tokens and Grouped Query Attention, it offers efficient inference and excels in technical and reasoning benchmarks for its size class.

Loading preview...