# R-4B: Incentivizing General-Purpose Auto-Thinking Capability in MLLMs via Bi-Mode Annealing and Reinforce Learning

## TL;DR

R-4B is an adaptive auto-thinking multimodal language model that selectively activates reasoning processes to improve efficiency and accuracy, outperforming existing models on diverse benchmarks.

## Contribution

The paper introduces R-4B, a novel auto-thinking MLLM with bi-mode annealing and policy optimization, enabling adaptive reasoning based on problem complexity.

## Key findings

- Achieves state-of-the-art results on 25 benchmarks.
- Outperforms Qwen2.5-VL-7B on most tasks.
- Comparable to larger models like Kimi-VL-A3B-Thinking-2506 (16B).

## Abstract

Multimodal Large Language Models (MLLMs) equipped with step-by-step thinking capabilities have demonstrated remarkable performance on complex reasoning problems. However, this thinking process is redundant for simple problems solvable without complex reasoning. To address this inefficiency, we propose R-4B, an auto-thinking MLLM, which can adaptively decide when to think based on problem complexity. The central idea of R-4B is to empower the model with both thinking and non-thinking capabilities using bi-mode annealing, and apply Bi-mode Policy Optimization (BPO) to improve the model's accuracy in determining whether to activate the thinking process. Specifically, we first train the model on a carefully curated dataset spanning various topics, which contains samples from both thinking and non-thinking modes. Then it undergoes a second phase of training under an improved GRPO framework, where the policy model is forced to generate responses from both modes for each input query. Experimental results show that R-4B achieves state-of-the-art performance across 25 challenging benchmarks. It outperforms Qwen2.5-VL-7B in most tasks and achieves performance comparable to larger models such as Kimi-VL-A3B-Thinking-2506 (16B) on reasoning-intensive benchmarks with lower computational cost.