WikiofAuto
WIKI INDEX/models/Qwen 3 & Qwen 3.8
modelsUpdated: 2026-11-01

Qwen 3 & Qwen 3.8

কিউয়েন ৩ এবং ৩.৮ (Qwen 3.8)

Qwen 3 and Qwen 3.8 are open-weight foundation, coding, and multimodal vision-language models (0.6B to 72B+ parameters) widely adopted for local GGUF inference, function calling, and QLoRA specialization.

Overview

The Qwen 3 / 3.8 model series dominates local AI and Kaggle competition leaderboards due to its dense and Mixture-of-Experts (MoE) variants, native ChatML tool-calling templates, and high-resolution mmproj vision projectors.