Qwen Releases Qwen3.8-27B With an Unchanged Architecture and Higher Agentic Scores

Sign in✓ Signed in
Qwen has released Qwen3.8-27B, a 27-billion-parameter open-weight vision-language model that retains the network architecture of Qwen3.6-27B while reporting broad gains in coding and agentic evaluations. The release suggests that its improvements come principally from training and post-training changes rather than a larger or redesigned model.
Continue reading
The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in✓ Signed in — this article isn’t included in your current plan.
Unlocking the full article...