vLLM ships DSpark draft model for Kimi K3.1 Pro with up to 30% higher acceptance length
vllm_project · x · 2026-11-18
vLLM, Inferact and Red Hat AI's Speculators team released the DSpark speculative-decoding draft model for Kimi K3.1 Pro: up to 30% higher acceptance length vs the preview on HumanEval, plus gains of 12-24% on long-context benchmarks. A 2K sliding window in every draft layer keeps draft KV memory bounded even at 1M-token context. Weights and vLLM configs are on Hugging Face.
More from Infra
- Inside Microsoft's massive Outlook outage: what broke and why fixes took so long — tomwarren · 2026-11-18
- Kimi K3.1 Pro DSpark model card: 2.5B draft, 5.57 acceptance length on GSM8K — vllm_project · 2026-11-18
- Wonderful runs in-house voice models on Together AI at 274ms first sentence — togethercompute · 2026-11-18
- Anthropic's S-1 Shows Only 25% of Compute Goes to Inference — Sparking a Bear Case — toptickcrypto · 2026-11-18
- LanceDB Raises $57.5M Series B for AI-Native Multimodal Data Stack — pzakin · 2026-11-18
- Multi-Branch Transformer scales LLMs via parallel branches, saving up to 1.26x training compute — KyeGomezB · 2026-11-18