Qwen is cooking hard
Summary
The author expresses anticipation for upcoming Qwen models: a 122B parameter version and a new 27B version.
Similar Articles
Can't wait to see Qwen3.8-27B
Qwen announced Qwen3.8, including a new 27B model, generating excitement for local deployment.
Bro wtf, Qwen Lab cooked with Qwen 3.8 27B, it's so fucking good
Qwen Lab has released Qwen 3.8 27B, which shows significant improvement over previous versions like Qwen 3.6 27B and other open-source models. The author hopes that Qwen publishes papers to help other labs develop similar high-quality small models.
Qwen will release another 27B with high probability
Qwen is highly likely to release a 27B parameter model, though the exact roadmap is still pending.
Qwen Developers' responses from their recent Twitter/X AMA
Qwen developers held an AMA on Twitter, confirming an upcoming 27B model, sharing details about Qwen 3.8's 2.4T parameters and 95B active, and noting the community's influence on releases.
Are we getting Qwen 3.8 35-A3B?
Speculation about the upcoming Qwen 3.8 release, questioning whether it will be a dense 27B model or a MoE variant like the previous 35B-A3B, with discussion of performance implications for local hardware.