ggml-org Releases Qwen3.8-27B-GGUF with Agent and Speculative Sampling
ggerganov · x · 2026-08-15
ggml-org released the Qwen3.8-27B-GGUF model in Q4 quantizations on DGX Spark. The demo command showcases features like speculative sampling (draft-mtp), default specs, and a reasoning-preserving agent mode.
Related event: Community Shares Qwen3.8-27B Deployment on 32GB VRAM(4 posts)→
More from Models
- Argon impresses at non-coding tasks, claimed to outscore all OpenAI and Claude models — bindureddy · 2026-10-17
- Huawei details training DeepSeek V4 on Ascend 950 supernodes: 273K tok/s on 2 racks — teortaxesTex · 2026-10-17
- Model 6.1 Sol Solves a 6D Rubik's Cube, Far Beyond Human Intuition — flowersslop · 2026-10-17
- Reddit user slams Codex for over-engineering: turning a 30-second task into an hour-long run — Reaper_1492 · 2026-10-17
- 10 open-source repos for training small models to do your agent's job — victor_explore · 2026-10-17
- Claude Sonnet 4.5 will be fully removed from AWS Bedrock by April 2027 — repligate · 2026-10-17