# CapRL: Stimulating Dense Image Caption Capabilities via Reinforcement Learning

arXiv: https://arxiv.org/abs/2509.22647 · HTML: https://mlsanity.com/paper/2509.22647/

## Key facts

- Cited by 38 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 57 of its 66 references that are in the map.
- First posted on arXiv 26 Sep 2025.

## Placed in

- LLM reasoning / [Multimodal reasoning](https://mlsanity.com/lr/multimodal-reasoning/)
- Multimodal LLMs / vision-language models / [VLM benchmarks and visual reasoning](https://mlsanity.com/mllm/vlm-benchmarks-and-visual-reasoning/)

## Builds on (most cited first, 30 of 57)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation](https://mlsanity.com/paper/2201.12086/) (2022)
- [LAION-5B: An open large-scale dataset for training next generation image-text models](https://mlsanity.com/paper/2210.08402/) (2022)
- [Conceptual 12M: Pushing Web-Scale Image-Text Pre-Training To Recognize Long-Tail Visual Concepts](https://mlsanity.com/paper/2102.08981/) (2021)
- [Kimi k1.5: Scaling Reinforcement Learning with LLMs](https://mlsanity.com/paper/2501.12599/) (2025)
- [Tulu 3: Pushing Frontiers in Open Language Model Post-Training](https://mlsanity.com/paper/2411.15124/) (2024)
- [ShareGPT4V: Improving Large Multi-Modal Models with Better Captions](https://mlsanity.com/paper/2311.12793/) (2023)
- [Deep Visual-Semantic Alignments for Generating Image Descriptions](https://mlsanity.com/paper/1412.2306/) (2014)
- [Large Language Diffusion Models](https://mlsanity.com/paper/2502.09992/) (2025)
- [Show and Tell: A Neural Image Caption Generator](https://mlsanity.com/paper/1411.4555/) (2014)
- [SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-training](https://mlsanity.com/paper/2501.17161/) (2025)
- [Visual-RFT: Visual Reinforcement Fine-Tuning](https://mlsanity.com/paper/2503.01785/) (2025)
- [TimeChat: A Time-sensitive Multimodal Large Language Model for Long Video Understanding](https://mlsanity.com/paper/2312.02051/) (2023)
- [ShareGPT4Video: Improving Video Understanding and Generation with Better Captions](https://mlsanity.com/paper/2406.04325/) (2024)
- [Rubrics as Rewards: Reinforcement Learning Beyond Verifiable Domains](https://mlsanity.com/paper/2507.17746/) (2025)
- [InternLM-XComposer2: Mastering Free-form Text-Image Composition and Comprehension in Vision-Language Large Model](https://mlsanity.com/paper/2401.16420/) (2024)
- [SemDeDup: Data-efficient learning at web-scale through semantic deduplication](https://mlsanity.com/paper/2303.09540/) (2023)
- [Long-CLIP: Unlocking the Long-Text Capability of CLIP](https://mlsanity.com/paper/2403.15378/) (2024)
- [PyramidDrop: Accelerating Your Large Vision-Language Models via Pyramid Visual Redundancy Reduction](https://mlsanity.com/paper/2410.17247/) (2024)
- [GUI-R1 : A Generalist R1-Style Vision-Language Action Model For GUI Agents](https://mlsanity.com/paper/2504.10458/) (2025)
- [Inference-Time Scaling for Generalist Reward Modeling](https://mlsanity.com/paper/2504.02495/) (2025)
- [Improving CLIP Training with Language Rewrites](https://mlsanity.com/paper/2305.20088/) (2023)
- [ALLaVA: Harnessing GPT4V-Synthesized Data for Lite Vision-Language Models](https://mlsanity.com/paper/2402.11684/) (2024)
- [Unified Reward Model for Multimodal Understanding and Generation](https://mlsanity.com/paper/2503.05236/) (2025)
- [Crossing the Reward Bridge: Expanding RL with Verifiable Rewards Across Diverse Domains](https://mlsanity.com/paper/2503.23829/) (2025)

## Built on by (most cited first, 30 of 38)

- [Spatial-SSRL: Enhancing Spatial Understanding via Self-Supervised Reinforcement Learning](https://mlsanity.com/paper/2510.27606/) (2025)
- [Intern-S1-Pro: Scientific Multimodal Foundation Model at Trillion Scale](https://mlsanity.com/paper/2603.25040/) (2026)
- [TimeChat-Captioner: Scripting Multi-Scene Videos with Time-Aware and Structural Audio-Visual Captions](https://mlsanity.com/paper/2602.08711/) (2026)
- [VideoChat3: Fully Open Video MLLM for Efficient and Generalist Video Understanding](https://mlsanity.com/paper/2607.14935/) (2026)
- [RubiCap: Rubric-Guided Reinforcement Learning for Dense Image Captioning](https://mlsanity.com/paper/2603.09160/) (2026)
- [Seeing with You: Perception-Reasoning Coevolution for Multimodal Reasoning](https://mlsanity.com/paper/2603.28618/) (2026)
- [CCCaption: Dual-Reward Reinforcement Learning for Complete and Correct Image Captioning](https://mlsanity.com/paper/2602.21655/) (2026)
- [CapRL++: Unified Reinforcement Learning with Verifiable Rewards for Dense Image and Video Captioning](https://mlsanity.com/paper/2606.09393/) (2026)
- [CapGeo-Bench: Decoupling Visual Perception from Reasoning and Evaluating Geometric Understanding](https://mlsanity.com/paper/2510.09302/) (2025)
- [Visual Self-Refine: A Pixel-Guided Paradigm for Accurate Chart Parsing](https://mlsanity.com/paper/2602.16455/) (2026)
- [TRivia: Self-supervised Fine-tuning of Vision-Language Models for Table Recognition](https://mlsanity.com/paper/2512.01248/) (2025)
- [Intern-S2-Preview: Scientific Agentic Foundation Model](https://mlsanity.com/paper/2608.13505/) (2026)
- [When Thinking Hurts: Mitigating Visual Forgetting in Video Reasoning via Frame Repetition](https://mlsanity.com/paper/2603.16256/) (2026)
- [VideoSeeker: Incentivizing Instance-level Video Understanding via Native Agentic Tool Invocation](https://mlsanity.com/paper/2605.16079/) (2026)
- [Decouple to Generalize: Context-First Self-Evolving Learning for Data-Scarce Vision-Language Reasoning](https://mlsanity.com/paper/2512.06835/) (2025)
- [CharTide: Data-Centric Chart-to-Code Generation via Tri-Perspective Tuning and Inquiry-Driven Evolution](https://mlsanity.com/paper/2604.22192/) (2026)
- [Let ViT Speak: Generative Language-Image Pre-training](https://mlsanity.com/paper/2605.00809/) (2026)
- [CineCap: Structured Reasoning with Spatio-Temporal Anchors for Cinematographic Video Captioning](https://mlsanity.com/paper/2606.24636/) (2026)
- [AVSCap: Orchestrating Audio-Visual Synergy for Omni-modal Video Captioning](https://mlsanity.com/paper/2607.12820/) (2026)
- [Grounded-Exo2Ego: Structured Semantic Grounding for Robust Exocentric-to-Egocentric Video Generation](https://mlsanity.com/paper/2608.20534/) (2026)
- [Momentum-Coupled Rubric Adaptation for Detailed Image Captioning](https://mlsanity.com/paper/2609.36893/) (2026)
- [DynaSolidGeo: A Dynamic Benchmark for Genuine Spatial Mathematical Reasoning of VLMs in Solid Geometry](https://mlsanity.com/paper/2510.22340/) (2025)
- [MVP: Enhancing Video Large Language Models via Self-supervised Masked Video Prediction](https://mlsanity.com/paper/2601.03781/) (2026)
- [Dynamic Token Compression for Efficient Video Understanding through Reinforcement Learning](https://mlsanity.com/paper/2603.26365/) (2026)
- [StableI2I: Spotting Unintended Changes in Image-to-Image Transition](https://mlsanity.com/paper/2605.04453/) (2026)
- [VCap: Hypergeometric Rewards for Weak-to-Strong Visual Captioning](https://mlsanity.com/paper/2605.28023/) (2026)
- [Actor as Its Own Critic: Unifying Region Understanding and Localization via CycleGRPO](https://mlsanity.com/paper/2607.11581/) (2026)
- [SciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding](https://mlsanity.com/paper/2609.05141/) (2026)
- [Cross-modal Identity Mapping: Minimizing Information Loss in Modality Conversion via Reinforcement Learning](https://mlsanity.com/paper/2603.01696/) (2026)
- [A Reconstruction-Based Framework for Caption Evaluation Beyond Reference Captions](https://mlsanity.com/paper/2607.23235/) (2026)

## Abstract (arXiv)

Image captioning is a fundamental task that bridges the visual and linguistic domains, playing a critical role in pre-training Large Vision-Language Models (LVLMs). Current state-of-the-art captioning models are typically trained with Supervised Fine-Tuning (SFT), a paradigm that relies on expensive, non-scalable data annotated by humans or proprietary models. This approach often leads to models that memorize specific ground-truth answers, limiting their generality and ability to generate diverse, creative descriptions. To overcome the limitation of SFT, we propose applying the Reinforcement Learning with Verifiable Rewards (RLVR) paradigm to the open-ended task of image captioning. A primary challenge, however, is designing an objective reward function for the inherently subjective nature of what constitutes a "good" caption. We introduce Captioning Reinforcement Learning (CapRL), a novel training framework that redefines caption quality through its utility: a high-quality caption should enable a non-visual language model to accurately answer questions about the corresponding image. CapRL employs a decoupled two-stage pipeline where an LVLM generates a caption, and the objective reward is derived from the accuracy of a separate, vision-free LLM answering Multiple-Choice Questions based solely on that caption. As the first study to apply RLVR to the subjective image captioning task, we demonstrate that CapRL significantly enhances multiple settings. Pretraining on the CapRL-5M caption dataset annotated by CapRL-3B results in substantial gains across 12 benchmarks. Moreover, within the Prism Framework for caption quality evaluation, CapRL achieves performance comparable to Qwen2.5-VL-72B, while exceeding the baseline by an average margin of 8.4%. Code is available here: https://github.com/InternLM/CapRL.