DFlash 2: Qwen3.8-27B hits 70 tok/s on Apple M5 Max
Lianhuiq · x · 2026-08-19
Inco AI released DFlash 2, an inference optimization technique. A demo shows Qwen3.8-27B running at 70 tok/s on an M5 Max MacBook Pro, which is 4.6× faster than autoregressive decoding.
Key Improvements:
- Over 20% more output per verification pass with 1% added latency.
- Predicts the entire token block in parallel instead of autoregressive drafting.
- Widely adopted in the ecosystem by NVIDIA, Google, and CoreWeave.
Related event: Inco AI Launches DFlash 2, Speeding Up LLM Inference Up to 4.6x(4 posts)→
More from Infra
- Phison CEO calls CXMT memory scare a 'ghost story', warns of materials shortage — BenBajarin · 2026-10-16
- Ben Bajarin: The Future of Compute Is Fungible as Hyperscalers Design for Workload Flexibility — BenBajarin · 2026-10-16
- AirLLM runs frontier-size models layer-by-layer: 671B on 12GB VRAM — Aiden_Tech_Ai · 2026-10-16
- NVIDIA Launches cuPotential in CUDA-X to Bring AI Agents to Semiconductor Materials Simulation — tomaszbednarz · 2026-10-16
- Q4-quantized coder model on 8-core CPU builds interactive solar system demo in one shot — IntrepidMindExplorer · 2026-10-16
- TokenSpeed MLA 2 ships for Vera Rubin: up to 3.44× faster Kimi K3 decode — zhyncs42 · 2026-10-16