NVIDIA Provides Day-0 Support for Alibaba's Qwen3.8-Flash-Next with NeMo

Alibaba_Qwen · x · 2026-08-27

NVIDIA announced Day-0 support for Alibaba's Qwen3.8-Flash-Next model. Developers can now use NVIDIA NeMo AutoModel and NeMo RL to fine-tune the model for domain-specific use cases. Recipes are also available to run the model with sgl, vllm, and TokenSpeed. The model is a hybrid-attention MoE architecture with 125B parameters.

Related event: Alibaba Open-Sources Qwen3.8-Flash-Next, A 6B-Activated Preview of Qwen4 Architecture(18 posts)→

Original post →

More from Infra

Infra channel →