Tencent open-sources AngelSpec, a speculative decoding framework with up to 2.4x speedup
teortaxesTex · x · 2026-07-29
Tencent has open-sourced AngelSpec, a Torch-native framework for training and deploying speculative decoding drafters.
- It supports both autoregressive MTP drafting and block-parallel DFlash-style decoding.
- On Hy3-A21B, the DFly family reportedly delivers 1.98–2.40× end-to-end speedups over autoregressive decoding across concurrency levels from 4 to 64.
- Tencent also released the training code, Hy3-A21B MTP/DFly drafter weights, a paper, docs, and Hugging Face / ModelScope resources.
- The repo positions AngelSpec as the framework behind the technical report’s released drafters.
Related event: Tencent Open-Sources AngelSpec Speculative Decoding Framework(4 posts)→
More from Infra
- Reflection AI open-sources Beam under Apache 2.0 with 40+ ecosystem partners — zhyncs42 · 2026-10-16
- NVIDIA launches RTX Spark for local AI on Windows PCs — ostrisai · 2026-10-16
- SkyRL ships Day-0 RL support for Beam with +62% LoRA decode throughput in vLLM — vllm_project · 2026-10-16
- DFlash2 delivers up to 3.03x decode-only speedup for Beam across five benchmarks — vllm_project · 2026-10-16
- vLLM's Beam serving kernels cut GB200 forward time from 8.15 to 6.47 ms at TP8 batch 128 — vllm_project · 2026-10-16
- Reflection's 501B MoE Beam hits 463 tok/s on 8x GB200 with NVFP4 at day 0 — zhyncs42 · 2026-10-16