Qwen vs. WizardLM

Overview

Qwen is Alibaba's broad family of open and hosted models for reasoning, coding, multimodal understanding, and agents. Qwen 3.8 Max Preview is the current generation.

Qwen, formerly Tongyi Qianwen, is a family of language and multimodal models developed by Alibaba. It includes general-purpose, coding, vision-language, audio, embedding, and reasoning variants. Qwen 3.8 Max Preview is the current hosted flagship, with support for multimodal and agentic workflows. Earlier Qwen generations, including Qwen 3.6 checkpoints, remain available as Apache 2.0 open weights; availability and licensing should therefore be checked per model.

Initial release: 2023-09-13

Current generation: Qwen 3.8

Reference

https://qwen.ai/

Further Reading

• https://arxiv.org/abs/2309.16609
• https://huggingface.co/Qwen
• https://qwen.ai/blog?id=qwen3.6-35b-a3b

Overview

WizardLM is a variant of LLaMA trained with complex instructions.

WizardLM is a LLM based on LLaMA trained using a new method, called Evol-Instruct, on complex instruction data. By using AI to "evolve" instructions, WizardLM outperforms similar LLaMA-based LLMs trained on simpler instruction data. On the 6th of July, 2023, WizardLM V1.1 was released with significantly improved performance, and as of 15 April 2024, WizardLM-2 was released with state-of-the-art performance.

Initial release: 2023-05-26

Reference

https://github.com/nlpxucan/WizardLM

Further Reading

• https://huggingface.co/WizardLM
• https://arxiv.org/abs/2304.12244
• https://www.reddit.com/r/LocalLLaMA/comments/142iw20/official_wizardlm30b_v10_released_can_beat/