# RelayGen: Intra-Generation Model Switching for Efficient Reasoning

arXiv: https://arxiv.org/abs/2602.06454 · HTML: https://mlsanity.com/paper/2602.06454/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 19 of its 27 references that are in the map.
- First posted on arXiv 6 Feb 2026.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 19 of 19)

- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Let's Verify Step by Step](https://mlsanity.com/paper/2305.20050/) (2023)
- [OpenAI o1 System Card](https://mlsanity.com/paper/2412.16720/) (2024)
- [FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performance](https://mlsanity.com/paper/2305.05176/) (2023)
- [Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads](https://mlsanity.com/paper/2401.10774/) (2024)
- [DeepSeek-V3 Technical Report](https://mlsanity.com/paper/2412.19437/) (2024)
- [RouteLLM: Learning to Route LLMs with Preference Data](https://mlsanity.com/paper/2406.18665/) (2024)
- [Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models](https://mlsanity.com/paper/2503.16419/) (2025)
- [EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Test](https://mlsanity.com/paper/2503.01840/) (2025)
- [Wait, We Don't Need to "Wait"! Removing Thinking Tokens Improves Reasoning Efficiency](https://mlsanity.com/paper/2506.08343/) (2025)
- [Efficient Inference for Large Reasoning Models: A Survey](https://mlsanity.com/paper/2503.23077/) (2025)
- [Speculative Thinking: Enhancing Small-Model Reasoning with Large Model Guidance at Inference Time](https://mlsanity.com/paper/2504.12329/) (2025)
- [The Avengers: A Simple Recipe for Uniting Smaller Language Models to Challenge Proprietary Giants](https://mlsanity.com/paper/2505.19797/) (2025)
- [R2R: Efficiently Navigating Divergent Reasoning Paths with Small-Large Model Token Routing](https://mlsanity.com/paper/2505.21600/) (2025)
- [Route to Reason: Adaptive Routing for LLM and Reasoning Strategy Selection](https://mlsanity.com/paper/2505.19435/) (2025)
- [CP-Router: An Uncertainty-Aware Router Between LLM and LRM](https://mlsanity.com/paper/2505.19970/) (2025)
- [SplitReason: Learning To Offload Reasoning](https://mlsanity.com/paper/2504.16379/) (2025)
- [R-Stitch: Dynamic Trajectory Stitching for Efficient Reasoning](https://mlsanity.com/paper/2507.17307/) (2025)

## Built on by (most cited first, 1 of 1)

- [What Does a Token Cost? A Mixture-of-Agents Measurement of Sufficient Per-Token Compute](https://mlsanity.com/paper/2610.02491/) (2026)

## Abstract (arXiv)

Large reasoning models (LRMs) achieve strong performance on complex reasoning tasks by generating long, multi-step reasoning trajectories, but inference-time scaling incurs substantial deployment cost. A key challenge is that generation difficulty varies within a single output, whereas existing efficiency-oriented approaches either ignore this intra-generation variation or rely on supervised token-level routing with high system complexity. We present RelayGen, a training-free, segment-level runtime model switching framework that exploits difficulty variation in long-form reasoning. Through offline analysis of generation uncertainty using token probability margins, we show that coarse-grained segment-level control is sufficient to capture difficulty transitions within a reasoning trajectory. RelayGen identifies model-specific switch cues that signal transitions to lower-difficulty segments and dynamically delegates their continuation to a smaller model, while preserving high-difficulty reasoning on the large model. Across multiple reasoning benchmarks, RelayGen substantially reduces inference latency while preserving most of the accuracy of large models. When combined with speculative decoding, RelayGen achieves up to 2.2× end-to-end speedup with less than 2% accuracy degradation, without requiring additional training or learned routing components.