New model from Telperion: Qwen3.8-27B-MLA & Qwen3.8-27B-GLA-g2. 4x smaller KV cache, smarter, and more efficient reasoning.

🌐 链接: https://www.reddit.com/r/LocalLLaMA/comments/1x5wcnz/new_model_from_telperion_qwen3827bmla/
🏷️ 分组: Reddit｜AI·自托管·VPS
🕒 时间: 2026-10-15 00:47:46
Reddit
From the LocalLLaMA community on Reddit: New model from Telperion: Qwen3.8-27B-MLA & Qwen3.8-27B-GLA-g2. 4x smaller KV cache, smarter…
From the LocalLLaMA community on Reddit: New model from Telperion: Qwen3.8-27B-MLA & Qwen3.8-27B-GLA-g2. 4x smaller KV cache, smarter…
Explore this post and more from the LocalLLaMA community