Qwen Releases Qwen3.8-27B: A 27B Multimodal AI Model with Vision and Coding Strengths
Qwen has released Qwen3.8-27B, a 27-billion-parameter language model built on the Qwen3.5 architecture that supports both text and visual inputs, including image and video understanding. The model features a native context window of 262,144 tokens, extendable to one million tokens, and was trained with multi-token prediction for faster inference. By default, it operates in a thinking mode that generates step-by-step reasoning chains before delivering final responses, though this can be turned off when cost efficiency is needed. On key benchmarks, it scores 61.7% on SWE-bench Pro for software engineering tasks and 84.3% on OSWorld-Verified for computer-use automation, outperforming several earlier models. Available via Hugging Face Transformers, it is also compatible with inference frameworks such as vLLM and SGLang.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Discussion (0)
Log in to join the discussion and vote.
Log in