Qwen Open Source Qwen3.8-27B: Only 17GB Needed for Local Quantitative Analysis

According to Perceive Beating monitoring, Qwen has officially released the Qwen3.8-27B weight. This is the version of the Qwen3.8 series specifically designed for on-premise deployment, utilizing the 27B dense architecture. The official model card indicates that it natively supports image and video understanding, with a context size of 262,144 Tokens, extendable to around 1 million Tokens.

Of greater importance is the hardware threshold. Unsloth has followed up with quantization support, stating that Qwen3.8-27B can run locally on approximately 17GB of RAM or VRAM after quantization. In other words, a 24GB consumer-grade GPU, or even certain high-memory Macs, can run Qwen3.8 on their own machines.

Previously, the 2.4T parameter Qwen3.8 large model weight that was released only supported text and required activation of the pensive mode; in contrast, the 27B version retains support for image and video input.

For most individual developers, the 27B version is the true usable open-source release of this round of Qwen3.8.