DFlash 2 runs a 27B model at 70 tok/s on M5 Max, up to 4.6x faster
DevDminGod · x · 2026-08-19
DFlash 2 pushes Qwen3.8-27B to 70 tok/s on an M5 Max MacBook Pro — up to 4.6× faster than standard autoregressive decoding while keeping the same output.
DFlash 2 is the next generation of the technique, seeded at Z Lab and upgraded by Inco AI. A 27B open-weight model is now genuinely fast enough for interactive local use — local AI is speeding up almost as fast as it's getting smarter.
Related event: Inco AI Launches DFlash 2, Speeding Up LLM Inference Up to 4.6x(4 posts)→
More from Infra
- Jensen Huang: Musk's AI buildout moved from K's to M's, now a CSP-tier NVIDIA customer — elonmusk · 2026-10-16
- EpochAI pegs GPT-6 training at up to 1.8e27 FLOP; scaling01 calls param estimates way too high — scaling01 · 2026-10-16
- Meta's Muse Spark Open scores 39 vs 48 for closed Spark 1.3 on AA Intelligence Index — EAccelerate_42 · 2026-10-16
- LLM spots useless 95GB-per-frame memcpy in ffmpeg dblur, cutting runtime from 2452ms to 49ms — IgorCarron · 2026-10-16
- DeepSeek-V4-Flash on 4x Radeon AI PRO R9700: the 'refused' 4-card config hits ~79 tok/s — brainchillzZ · 2026-10-16
- Tree packing lands in TRL's AsyncGRPO, cutting agentic RL training compute by over 26x — joecole · 2026-10-16