# OmniDraft: A Cross-vocabulary, Online Adaptive Drafter for On-device Speculative Decoding

arXiv: https://arxiv.org/abs/2507.02659 · HTML: https://mlsanity.com/paper/2507.02659/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in Speculative decoding, one of 12 directions in Efficient LLM inference and serving.
- Builds on 47 of its 52 references that are in the map.
- First posted on arXiv 3 Jul 2025.

## Placed in

- Efficient LLM inference and serving / [Speculative decoding](https://mlsanity.com/inference/speculative-decoding/)

## Builds on (most cited first, 30 of 47)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Qwen2.5 Technical Report](https://mlsanity.com/paper/2412.15115/) (2024)
- [Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks](https://mlsanity.com/paper/1703.03400/) (2017)
- [Program Synthesis with Large Language Models](https://mlsanity.com/paper/2108.07732/) (2021)
- [Prototypical Networks for Few-shot Learning](https://mlsanity.com/paper/1703.05175/) (2017)
- [SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing](https://mlsanity.com/paper/1808.06226/) (2018)
- [Fast Inference from Transformers via Speculative Decoding](https://mlsanity.com/paper/2211.17192/) (2022)
- [Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation](https://mlsanity.com/paper/1609.08144/) (2016)
- [Don't Give Me the Details, Just the Summary! Topic-Aware Convolutional Neural Networks for Extreme Summarization](https://mlsanity.com/paper/1808.08745/) (2018)
- [Accelerating Large Language Model Decoding with Speculative Sampling](https://mlsanity.com/paper/2302.01318/) (2023)
- [On-Policy Distillation of Language Models: Learning from Self-Generated Mistakes](https://mlsanity.com/paper/2306.13649/) (2023)
- [Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads](https://mlsanity.com/paper/2401.10774/) (2024)
- [Sequence-Level Knowledge Distillation](https://mlsanity.com/paper/1606.07947/) (2016)
- [EAGLE: Speculative Sampling Requires Rethinking Feature Uncertainty](https://mlsanity.com/paper/2401.15077/) (2024)
- [RL2: Fast Reinforcement Learning via Slow Reinforcement Learning](https://mlsanity.com/paper/1611.02779/) (2016)
- [Efficient Off-Policy Meta-Reinforcement Learning via Probabilistic Context Variables](https://mlsanity.com/paper/1903.08254/) (2019)
- [EAGLE-2: Faster Inference of Language Models with Dynamic Draft Trees](https://mlsanity.com/paper/2406.16858/) (2024)
- [LayerSkip: Enabling Early Exit Inference and Self-Speculative Decoding](https://mlsanity.com/paper/2404.16710/) (2024)
- [Break the Sequential Dependency of LLM Inference Using Lookahead Decoding](https://mlsanity.com/paper/2402.02057/) (2024)
- [Byte Pair Encoding is Suboptimal for Language Model Pretraining](https://mlsanity.com/paper/2004.03720/) (2020)
- [DistillSpec: Improving Speculative Decoding via Knowledge Distillation](https://mlsanity.com/paper/2310.08461/) (2023)
- [Speculative Decoding with Big Little Decoder](https://mlsanity.com/paper/2302.07863/) (2023)
- [REST: Retrieval-Based Speculative Decoding](https://mlsanity.com/paper/2311.08252/) (2023)
- [f-Divergence Minimization for Sequence-Level Knowledge Distillation](https://mlsanity.com/paper/2307.15190/) (2023)
- [Online Speculative Decoding](https://mlsanity.com/paper/2310.07177/) (2023)
- [Why Exposure Bias Matters: An Imitation Learning Perspective of Error Accumulation in Language Generation](https://mlsanity.com/paper/2204.01171/) (2022)
- [Sequoia: Scalable, Robust, and Hardware-aware Speculative Decoding](https://mlsanity.com/paper/2402.12374/) (2024)

## Built on by (most cited first, 4 of 4)

- [TIDE: Temporal Incremental Draft Engine for Self-Improving LLM Inference](https://mlsanity.com/paper/2602.05145/) (2026)
- [Cross-Family Speculative Decoding for Polish Language Models on Apple~Silicon: An Empirical Evaluation of Bielik~11B with UAG-Extended MLX-LM](https://mlsanity.com/paper/2604.16368/) (2026)
- [X-CoSD: Communication-Efficient Cross-Vocabulary Collaborative Speculative Decoding](https://mlsanity.com/paper/2609.09166/) (2026)
- [Osprey: Target-agnostic Pre-training Makes Stronger Drafters in Speculative Decoding](https://mlsanity.com/paper/2609.09338/) (2026)

## Abstract (arXiv)

Speculative decoding generally dictates having a small, efficient draft model that is either pretrained or distilled offline to a particular target model series, for instance, Llama or Qwen models. However, within online deployment settings, there are two major challenges: 1) usage of a target model that is incompatible with the draft model; 2) expectation of latency improvements over usage and time. In this work, we propose OmniDraft, a unified framework that enables a single draft model to operate with any target model and adapt dynamically to user data. We introduce an online n-gram cache with hybrid distillation fine-tuning to address the cross-vocabulary mismatch across draft and target models; and further improve decoding speed by leveraging adaptive drafting techniques. OmniDraft is particularly suitable for on-device LLM applications where model cost, efficiency and user customization are the major points of contention. This further highlights the need to tackle the above challenges and motivates the textit``one drafter for all'' paradigm. We showcase the proficiency of the OmniDraft framework by performing online learning on math reasoning, coding and text generation tasks. Notably, OmniDraft enables a single Llama-68M model to pair with various target models including Vicuna-7B, Qwen2-7B and Llama3-8B models for speculative decoding; and additionally provides up to 1.5-2x speedup.