SanjiWatsuki/Sonya-7B
TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kPublished:Dec 31, 2023License:cc-by-4.0Architecture:Transformer0.0K Open Weights Cold

SanjiWatsuki/Sonya-7B is a 7 billion parameter merged language model based on the Mistral-7B-v0.1 architecture, designed for general-purpose conversational tasks including assistant and roleplay applications. This model distinguishes itself by achieving top performance on the MT-Bench benchmark, notably ranking #1 in the first turn and #2 overall, surpassing larger models like GPT-3.5 and Llama-2-70b-chat. It is optimized for a standard 4096 token context window, with experimental support for up to 16384 tokens using NTK scaling.

Loading preview...