Qwen3.8-27B

We promised unfastened weights for Qwen3.8. Now, clip to meet them! 🎉 ⚡ Qwen3.8-27B: - A autochthonal multimodal dense model. With conscionable 27B parameters, it outperforms Qwen3.7-Plus wide and shines successful real-world coding & agency workflows. - 262K autochthonal context, easy extendable to 1M tokens via YaRN. - Built for builders. Highly efficient, high-quality, and licensed nether Apache 2.0. 🚀 The unfastened weights for Qwen3.8-2.4T-A95B (Max-level) person besides been released recently. Whether you're shipping lightweight applications pinch Qwen3.8-27B locally aliases building agents pinch Qwen3.8-2.4T-A95B, they're yours now! Download, deploy, and build thing we haven't imagined yet. 👀👇 - Hugging Face： huggingface.co/collections/Qw… - ModelScope： modelscope.cn/collections/Qw…