Demystifying Hybrid Thinking: Can LLMs Truly Switch Between Think and No-Think?

TL;DR

Demystifying Hybrid Thinking investigates whether LLMs can truly switch between thinking and no-thinking. The authors demonstrate that current hybrid thinking models exhibit leakage of reasoning into the no-think mode, signaling limited controllability. They systematically identify data scale, paired versus non-paired data, no-think data proportion, and two-phase training as the four most influential factors, validating these across benchmarks like MATH500, AIME24, and GPQA. A practical recipe—emphasizing more no-think data and a two-phase training regimen—achieves comparable think-mode performance while substantially reducing no-think verbosity and reasoning traces, offering a path toward more controllable and efficient hybrid reasoning.

Abstract

Hybrid thinking enables LLMs to switch between reasoning and direct answering, offering a balance between efficiency and reasoning capability. Yet our experiments reveal that current hybrid thinking LLMs only achieve partial mode separation: reasoning behaviors often leak into the no-think mode. To understand and mitigate this, we analyze the factors influencing controllability and identify four that matter most: (1) larger data scale, (2) using think and no-think answers from different questions rather than the same question, (3) a moderate increase in no-think data number, and (4) a two-phase strategy that first trains reasoning ability and then applies hybrid think training. Building on these findings, we propose a practical recipe that, compared to standard training, can maintain accuracy in both modes while significantly reducing no-think output length (from $1085$ to $585$ on MATH500) and occurrences of reasoning-supportive tokens such as ``\texttt{wait}'' (from $5917$ to $522$ on MATH500). Our findings highlight the limitations of current hybrid thinking and offer directions for strengthening its controllability.

Figures (6)

• Figure 1: Case study on an AIME24 problem. We compare the responses of Qwen3-8B under no-think and think modes. In the no-think mode, Qwen3-8B still performs reasoning outside the no-think constraint (e.g., generating reasoning-supportive words such as Wait), indicating that its hybrid thinking ability remains imperfect and cannot achieve full control.
• Figure 2: Comparison of the performance of no-think mode of Hybrid model (Qwen3-8B) and Instruct model (Qwen2.5-7B-Instruct) on MATH500, AIME24, and GPQA. We show bar plots of accuracy, average output length and wait count across these three datasets.
• Figure 3: Line plots of Qwen2.5-7B-Instruct performance with different training data scales (20k/40k/80k/140k) on MATH500, AIME24, and GPQA. We report accuracy (top) and average output length (bottom) under both think and no-think modes. The results show that while accuracy and output length in the think mode, as well as accuracy in the no-think mode, remain relatively stable across scales, the no-think output length decreases substantially with larger training data.
• Figure 4: Comparison of accuracy and output lengths under think and no-think modes on MATH500. The bar charts illustrate model behavior under different training settings (pairs vs. no-pairs), highlighting how no-pairs could reduce output length in no think mode.
• Figure 5: Comparison of output lengths under think and no-think modes on MATH500. The bar charts illustrate model behavior under different think-to-no-think data ratios, showing that increasing the proportion of no-think data reduces output length in the no-think mode.