Speculative decoding drafter benchmark: AngelSpec DFly beats EAGLE-3 variants
maxsloef · x · 2026-11-12
Updated artifact and repo benchmarking speculative-decoding drafters: with thinking off, AngelSpec DFly leads with 0.821 top-1 first-token accuracy and 3.46 mean correct tokens (of 7), ahead of its MTP block (0.779/2.79) and EAGLE-3 variants WhiFlash (0.750/2.47), Tengyunw (0.716/2.04), RedHatAI (0.671/1.63), AngelSlim (0.579/1.23). With thinking on, gains shrink but DFly still tops (0.769/2.56). Notes one DFly variant won't work for certain lenses.
Related event: Benchmark Compares Draft Models for Qwen3-8B Speculative Decoding(2 posts)→
More from Infra
- NVIDIA plunges 9% after hours on report board may fire Jensen Huang — djcows · 2026-11-12
- Vmax teases Campaign: self-play RL task generation taught agents to use Modal SDK — soleio · 2026-11-12
- JITEngine synthesizes custom LLM inference engines for Apple Silicon, up to 2.7x faster — CShorten30 · 2026-11-12
- LeapQuant: training-free quantization makes linear attention up to 4.25x faster — tianyin_xu · 2026-11-12
- Microsoft's MAI Transcribe 2 sees 12x usage growth, hitting 8.5M weekly transcriptions — mustafasuleyman · 2026-11-12
- gVisor joins CNCF with Modal as co-maintainer after upstreaming GPU and training support — josh_wills · 2026-11-12