Adaptive Look, Flexible Thought: Learning How and How Long to Think in Multimodal Reasoning

Quanlin Chen; Xiuwei Chen; Kun Xiang; Zehua Ma; Chunjin Ye; Boqiang Guo; Xianming Liu; Xiaodan Liang

cs.CVOct 11, 2026

Adaptive Look, Flexible Thought: Learning How and How Long to Think in Multimodal Reasoning

Authors: Quanlin Chen, Xiuwei Chen, Kun Xiang, Zehua Ma, Chunjin Ye, Boqiang Guo, Xianming Liu, Xiaodan Liang

Organizations: Sun Yat-sen University · Shenzhen Loop Area Institute · Beijing Zhongguancun Academy · Huawei Technologies Co., Ltd · Harbin Institute of Technology

Abstract

The introduction of long chain-of-thought (CoT) reasoning and external tools has substantially improved the performance of multimodal large language models (MLLMs) on general and reasoning tasks. However, existing fast-slow reasoning paradigms mainly regulate textual reasoning length, while visual computation introduced by external tools or visual tokens is often treated as an auxiliary component. This separation prevents MLLMs from jointly optimizing visual perception and textual reasoning. To address this issue, we propose L2T (adaptive Look, flexible Thought), a unified reasoning framework that jointly optimizes the form and depth of computation. L2T first discretizes tool-produced visual states into visual tokens, bringing visual perception and textual reasoning into a shared autoregressive space. Based on this unified representation, L2T introduces a modality-aware routing reward to select appropriate reasoning modes and a difficulty-aware fast-slow strategy to adapt reasoning depth according to task complexity. Experiments on general perception and reasoning-intensive tasks demonstrate the effectiveness of L2T. Compared with explicit tool-calling approaches, L2T maintains competitive performance while achieving 1.7--3.6 times higher throughput.

Figures & tables

Figure 1: Overview of modality-aware routing and difficulty-aware fast-slow optimization.

Panel of Figure 2 — Figure 2: Throughput and generation length comparison.

Open-Source Models
Models	Visual-Centric Perception				Reasoning-Centric Understanding
Models	V ∗	HR4K	HR8K	MME-Lite	MMMU-Pro	MMMU-Pro-V	MathVista	MMVet	AVG
Qwen2.5-VL-3B ( Qwen Team, 2025 )	74.9 †	66.1 †	63.5 †	42.4	31.6	28.5 †	62.3	61.8	53.9
Qwen2.5-VL-7B ( Qwen Team, 2025 )	76.4 †	68.4 †	64.6 †	44.9	38.3	35.6 †	68.2	67.1	57.9
Textual Reasoning Models
Vision-R1-7B ( Huang et al., 2026 )	80.6 †	75.6 †	70.5 †	51.5 †	38.5 †	35.3 †	72.1	76.4 †	62.6

Table 1: Performance comparison across perception and reasoningcentric multimodal benchmarks. The best result is shown in bold , and the second-best result is underlined . For benchmark results reported in prior papers, we directly cite the published numbers.For missing entries, we evaluate the corresponding models under the same environment and mark these results with † .

Ration of Visual Token
Category	standard (%)	w/o route reward (%)	w/o fast-slow
Visual-Centric	31.8	21.9 (-9.9)	–
Reasoning-Centric	11.56	11.03 (-0.53)	–
Average Length of CoT
Visual-Centric	206.8	292.6 (+85.8)	234.7 (+27.9)
Reasoning-Centric	416.2	547.0 (+130.8)	469.9 (+53.7)

Table 2: Ablation on Route Reward and Difficulty-aware fast-slow.

Components	Visual-Centric	Reasoning-Centric	Total $\Delta$
w/o visual token	70.6	55.4	-1.7
w/o two-stage rl	69.3	56.1	-2.3
standard	71.2	56.5	–

Table 3: Ablation on visual token and two-stage rl.

Appendix figures & tables14 assets

Supplementary material from the paper’s appendix.

Appendix

Models	SeePhys		SeePhysPro		AVG
Models	Vanilla	Vision-only	Level 2	Level 1	AVG
Qwen2.5-VL-7B	11.7	8.5	11.5	12.0	10.9
FAST-7B	13.4	10.1	11.5	13.0	12.0
PAPO-7B	11.6	9.7	9.0	12.0	10.6
Thyme-RL	10.7	7.6	11.0	11.5	10.2
Monet	11.1	5.8	15.0	12.5	11.1

Appendix

Table 4: Performance on SeePhys-family physics reasoning benchmarks.

Figure 3: Pareto analysis of our method and slow-thinking RLVR Baseline.

Figure 4: VQ-VT converts tool-produced intermediate visual states into discrete visual-token sequences, enabling visual and textual reasoning to share a unified autoregressive token space for adaptive reasoning depth.

Image condition	Similarity
Repeated original	90.2%
Same image, 2 $\times$ upscale	84.0%
stretch	83.0%
Horizontal flip	72.8%
Noise + small mask	69.6%
Paired HRBench8K image	67.6%

Appendix

Table 5: VQ code similarity to the original 4K image.

Components	Visual-Centric	Reasoning-Centric	Total $\Delta$
w/o threshold	70.5	56.0	-1.2
standard	71.2	56.5	–

Appendix

Table 6: Ablation on thresholds.

Panel of Figure 5 — Figure 5: Difficulty-dependent reasoning length and benchmark-level CoT statistics.

Category	standard (%)	overtraining (%)	$\Delta$
Average Length of CoT
Visual-Centric	206.8	10	+196.8
Reasoning-Centric	416.2	10	+406.2
Accuracy
Visual-Centric	71.2	68.6	+2.6
Reasoning-Centric	56 .5	44.3	+12.2

Appendix

Table 7: Ablation on Route Reward

Outcome / failure mode	$n$	Mean	Median
Correct-all	4,573	277.3	214
Correct-Perception	2,370	200.6	198
Correct-Reasoning	2,203	359.9	260
Perception failure	1,758	241.9	216
Knowledge failure	517	430.0	281
Reasoning-process failure	724	505.8	304.5

Appendix

Table 8: Mean and median CoT length for each answer outcome and failure mode.

Panel of Figure 6 — Figure 6: Successful case study of our method.

Panel of Figure 7 — Figure 7: Successful case study of our method.

Panel of Figure 8 — Figure 8: Failed case study of our method.

Panel of Figure 9 — Figure 9: Case study on Collapsed CoT.

Panel of Figure 10 — Figure 10: Pie chart of the datasets.

Panel of Figure 11 — Figure 11: Remaining ratio of informative rollout groups after online filtering during RL training.

Explore similar work

Balancing Efficiency and Efficacy: Training-Free Attention-Guided Switching Between Explicit and Latent Thoughts for MLLMs
Aug 4, 2026Haoqian Kang, Liupeng Li, Kuofeng Gao +5Efficient Multimodal Inference Multimodal Large Language Models
Harbin Institute of Technology, Shenzhen Shenzhen, China · Tsinghua Shenzhen International Graduate School, Tsinghua University Shenzhen, China · Peng Cheng Laboratory Shenzhen, China +1
CoLT: Teaching Multi-Modal Models to Think with Chain of Latent Thoughts
Jun 30, 2026Lianyu Hu, Shengqian Qin, Zeqin Liao +4Efficient Multimodal Inference Efficient Language Model Reasoning
Nanyang Technological University, Singapore · Shanghai Jiao Tong University, China · NKIARI, Shenzhen Futian, China +2
TVI-CoT: Text-Visual Interleaved Chain-of-Thought Reasoning for Multimodal Understanding
Jun 7, 2026Lianyu Hu, Xiaoyu Ma, Zeqin Liao +1Multimodal Large Language Models Multimodal Reasoning
College of Computing and Data Science of Nanyang Technological University, Singapore.

Aug 4, 2026J/K move · Enter open · S save

Balancing Efficiency and Efficacy: Training-Free Attention-Guided Switching Between Explicit and Latent Thoughts for MLLMs

Haoqian Kang, Liupeng Li, Kuofeng Gao, Jinpeng Wang+4

Harbin Institute of Technology, Shenzhen Shenzhen, China · Tsinghua Shenzhen International Graduate School, Tsinghua University Shenzhen, China · Peng Cheng Laboratory Shenzhen, China+1

Efficient Multimodal Inference Multimodal Large Language Models Latent Visual Reasoning VLM Reasoning

Open paper AskPDF arXiv

Reasoning in Multimodal Large Language Models (MLLMs) requires both fine-grained visual perception and rigorous logical deduction. Explicit text-based Chain-of-Thought (CoT) is computationally expensive and prone to visual hallucinations, while existing latent reasoning methods typically require costly training. Furthermore, directly adapting training-free LLM reasoning mechanisms to the multimodal setting yields unstable performance. We identify that this failure stems from their reliance on token-level entropy, which fundamentally conflates perceptual ambiguity (e.g., unclear visual details) with logical uncertainty (e.g., complex reasoning steps). To overcome this bottleneck, we present a novel training-free inference strategy for MLLMs that explicitly decouples perception and reasoning. We propose a novel metric, the vision-to-text attention ratio, to dynamically gauge the model's cognitive focus. Guided by this metric, our proposed framework, Attention-Guided Switching (AGS), adaptively triggers latent reasoning for perceptual tokens to preserve high-fidelity visual information in the continuous space, while enforcing explicit text generation for logical tokens to maintain structural anchoring. Extensive experiments demonstrate that our method achieves state-of-the-art performance, significantly improving both accuracy and inference efficiency by reducing autoregressive steps and latency. Code is released at https://github.com/swordAndSnow/MM26-AGS.

Abstract

Explore similar work

Balancing Efficiency and Efficacy: Training-Free Attention-Guided Switching Between Explicit and Latent Thoughts for MLLMs

CoLT: Teaching Multi-Modal Models to Think with Chain of Latent Thoughts

TVI-CoT: Text-Visual Interleaved Chain-of-Thought Reasoning for Multimodal Understanding