Alibaba releases Qwen3.8-27B under Apache-2.0 as a 27B dense multimodal model, with all card benchmarks self-reported and unverified

Sources: primary, quantizations, discussion
Summary: Alibaba published Qwen3.8-27B on Hugging Face under Apache-2.0, a 27B dense multimodal model. The model card reports 61.7 on SWE-Bench Pro. Every figure on the card is self-reported with no independent reproduction. GGUF quantizations were published the same day.
Why it matters: A 27B dense model under Apache-2.0 with weights and GGUF quantizations on the release date puts the claimed SWE-Bench Pro score inside hardware a single developer already owns, with no license negotiation and no API dependency.