# Mid-Think: Training-Free Intermediate-Budget Reasoning via Token-Level Triggers Source: [Mid-Think: Training-Free Intermediate-Budget Reasoning via Token-Level Triggers]() Collection index: [llms.txt]() Data updated: 2026-06-05T05:00:04.417Z Authors: Wang Yang, Debargha Ganguly, Xinpeng Li, Chaoda Song, Shouren Wang, Vikash Singh, Vipin Chaudhary, Xiaotian Han Published: 2026-01-11T19:19:39.000Z ## Abstract Hybrid reasoning language models are commonly controlled through high-level Think/No-think instructions to regulate reasoning behavior, yet we found that such mode switching is largely driven by a small set of trigger tokens rather than the instructions themselves. Through attention analysis and controlled prompting experiments, we show that a leading \`\`Okay'' token induces reasoning behavior, while the newline pattern following \`\`\'' suppresses it. Based on this observation, we propose Mid-Think, a simple training-free prompting format that combines these triggers to achieve intermediate-budget reasoning, consistently outperforming fixed-token and prompt-based baselines in terms of the accuracy-length trade-off. Furthermore, applying Mid-Think to RL training after SFT reduces training time by approximately 15% while improving final performance of Qwen3-8B on AIME from 69.8% to 72.4% and on GPQA from 58.5% to 61.1%, demonstrating its effectiveness for both inference-time control and RL-based reasoning training. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Mathematical Reasoning on AIME No-think (test)](): Mathematical Reasoning; AIME No-think (test); Accuracy: 62.9. - [Mathematical Reasoning on AIME Think (test)](): Mathematical Reasoning; AIME Think (test); Accuracy: 72.4. - [Question Answering on GPQA No-think (test)](): Question Answering; GPQA No-think (test); Accuracy: 47.8. - [Question Answering on GPQA Think (test)](): Question Answering; GPQA Think (test); Accuracy: 61.1. - [Mathematical Reasoning on MATH500 (test)](): Mathematical Reasoning; MATH500 (test); Accuracy: 94.1. ## Original sources and implementation [Original paper]() [PDF]()