# Gen-Searcher: Reinforcing Agentic Search for Image Generation

> Gen-Searcher is the first trained multimodal agent that uses reinforcement learning to perform multi-hop web searches for knowledge-intensive image generation, significantly improving performance on benchmarks.

- **Source:** [arXiv](https://arxiv.org/abs/2603.28767)
- **Published:** 2026-04-01
- **Permalink:** https://picx.dev/p/6nnvNK
- **Whiteboard:** https://picx.dev/p/6nnvNK/image

## Summary

# Gen-Searcher: Reinforcing Agentic Search for Image Generation - Summary

## Summary (Overview)
*   **First Multimodal Deep Search Agent for Image Generation:** Gen-Searcher is the first trained agent that performs multi-hop web search and reasoning to gather textual knowledge and visual references for knowledge-intensive image generation.
*   **Novel Data Pipeline and Benchmarks:** The authors constructed two high-quality training datasets (**Gen-Searcher-SFT-10k** and **Gen-Searcher-RL-6k**) and introduced the **KnowGen** benchmark for evaluating search-grounded generation, along with the **K-Score** metric.
*   **Dual Reward Reinforcement Learning:** The model is trained via a two-stage process (Supervised Fine-Tuning followed by Agentic Reinforcement Learning) using a novel **dual reward feedback** design that combines text-based and image-based rewards to provide stable and informative learning signals for GRPO training.
*   **Significant Performance Gains:** Gen-Searcher substantially improves image generation on knowledge-intensive tasks, boosting **Qwen-Image by ~16 points on KnowGen** and **~15 points on WISE**, and shows strong transferability across different image generators (e.g., Seedream 4.5, Nano Banana Pro).
*   **Open Foundation:** The project is fully open-sourced (data, models, code) to serve as a foundation for future research on search agents for image generation.

## Introduction and Theoretical Foundation
Recent text-to-image models are constrained by frozen internal knowledge, struggling with real-world prompts that require up-to-date or knowledge-intensive information (e.g., specific landmarks, new products). While some proprietary models support text search, they lack visual reference retrieval. Prior RAG-based methods are limited by static databases and shallow retrieval, and prompt-based workflows are brittle and suboptimal.

This paper introduces **Gen-Searcher**, the first attempt to train a multimodal deep search agent for image generation using **agentic reinforcement learning (RL)**. The core idea is to train an agent that can actively perform multi-hop web search, browse, and reason to gather both textual evidence and visual references, which are then used to create a grounded prompt for a downstream image generator. This addresses the fundamental limitation of frozen knowledge in generative models.

## Methodology

### 1. Dataset Construction Pipeline
A four-stage pipeline was created to generate training data, which did not naturally exist.
1.  **Text Prompt Construction:** Two strategies were used:
    *   **Primary:** Prompt engineering with **Gemini 3 Pro** to generate multi-hop search-intensive prompts across ~20 diverse categories (Anime, Celebrities, Physics, Art, etc.).
    *   **Complementary:** Converting existing deep research QA datasets into image-generation-oriented prompts, primarily for General News.
2.  **Agentic Trajectory Generation:** **Gemini 3 Pro** was used with search tools (`search`, `image_search`, `browse`) in a multi-turn loop to generate search trajectories, resulting in a final grounded prompt and selected reference images.
3.  **Ground-Truth Image Synthesis:** The final prompts were fed into **Nano Banana Pro** to synthesize corresponding images as ground truth.
4.  **Data Filtering & Curation:** **Seed1.8** was used to score and filter samples based on faithfulness, correctness, aesthetics, safety, etc., combined with rule-based filtering. This yielded ~17K high-quality samples.
    *   **Gen-Searcher-SFT-10k:** For supervised fine-tuning.
    *   **Gen-Searcher-RL 6k:** For reinforcement learning.
    *   **KnowGen Benchmark:** 630 human-verified, held-out evaluation samples.

### 2. KnowGen Benchmark & K-Score
**KnowGen** is a comprehensive benchmark for evaluating search-grounded image generation in real-world, knowledge-intensive scenarios.
*   **Categories:** Divided into two subsets:
    *   **Science & Knowledge:** Astronomy, Biology, Chemistry, Physics, Engineering, Medicine, Industry, Architecture, History, Geography, Religion, Politics, Culture, Art, Sports.
    *   **Pop Culture & News:** Anime, Games, Films, Celebrities, Posters, General News.
*   **Evaluation Metric - K-Score:** Uses GPT-4.1 as a judge to evaluate generated images from four dimensions, each scored on a scale of {0, 0.5, 1}:
    *   **Faithfulness:** Scene-structure level adherence to the prompt.
    *   **Visual Correctness:** Accuracy of grounded visual attributes vs. reference.
    *   **Text Accuracy:** Presence, legibility, and correctness of required readable text.
    *   **Aesthetics:** Overall visual quality and appeal.
*   The final **K-Score** is a weighted combination:
    $$ \text{K-Score} = 0.1 \cdot \text{Faithfulness} + 0.4 \cdot \text{Visual Correctness} + 0.4 \cdot \text{Text Accuracy} + 0.1 \cdot \text{Aesthetics} $$

### 3. Training Scheme
Gen-Searcher is initialized from **Qwen3-VL-8B-Instruct** and trained in two stages.
*   **Search Tools:** The agent is equipped with three tools: `search` (web text search), `image_search` (retrieve images via text query), and `browse` (analyze webpage content).
*   **Stage 1: Supervised Fine-Tuning (SFT):** Trained on **Gen-Searcher-SFT-10k** to learn basic multi-turn tool use for search, reasoning, and prompt composition.
*   **Stage 2: Agentic Reinforcement Learning (RL):** Trained on **Gen-Searcher-RL-6k** using **GRPO** to optimize search trajectories.
*   **Dual Reward Feedback Design:** To address the noise and instability of pure image-based rewards (due to generator variance), a combined reward is used:
    $$ R = (1 - \alpha) R_{\text{image}} + \alpha R_{\text{text}} \quad \text{(Equation 1)} $$
    *   $R_{\text{image}}$: **K-Score** of the final generated image.
    *   $R_{\text{text}}$: Text-based reward (scored 0-1 by GPT-4.1) evaluating if the gathered prompt contains sufficient/correct information for generation.
    *   $\alpha$: Balancing hyperparameter (set to 0.5).
*   **Optimization:** The policy is optimized using **GRPO**. For each sampled output $o_i$ under query $q$, the advantage is computed:
    $$ A_i = \frac{R_i - \text{mean}(\{R_j\})}{\text{std}(\{R_j\})} \quad \text{(Equation 2)} $$
    The final policy update follows the standard GRPO objective $J_{\text{GRPO}}$ (Equation 3 in the paper), which includes a clipped probability ratio and a KL divergence penalty.

## Empirical Validation / Results

### Main Results on KnowGen Benchmark

**Table 1: Performance of different models on the KnowGen benchmark.**
| Models | Science & Knowledge | Pop Culture & News | **Overall K-Score** |
| :--- | :--- | :--- | :--- |
| | Visual cor. | Text acc. | Visual cor. | Text acc. | |
| GPT-Image-1.5 | 29.25 | 40.14 | 29.43 | 46.22 | **44.97** |
| Nano Banana Pro | 39.46 | 49.32 | 30.51 | 53.37 | **50.38** |
| Seedream 4.5 | 14.46 | 26.19 | 12.50 | 31.77 | **31.01** |
| Qwen-Image | 6.80 | 0.34 | 7.59 | 1.40 | **14.98** |
| **Gen-Searcher-8B + Qwen-Image** | 26.87 | 17.18 | 25.30 | 23.55 | **31.52** |
| **Gen-Searcher-8B + Seedream 4.5** | 36.35 | 43.52 | 39.04 | 45.86 | **47.29** |
| **Gen-Searcher-8B + Nano Banana Pro** | 45.07 | 49.32 | 43.01 | 52.30 | **53.30** |

*   **KnowGen is Challenging:** Open-source models (Qwen-Image, FLUX, Z-Image) score only 9-15, showing the difficulty of knowledge-intensive generation.
*   **Effectiveness of Gen-Searcher:** Brings substantial gains across backbones.
    *   Improves **Qwen-Image from 14.98 to 31.52** (+16.54 points).
    *   Transfers effectively to other generators: improves **Seedream 4.5 from 31.01 to 47.29** (+16.28 points) and **Nano Banana Pro from 50.38 to 53.30**.
*   **Dimension Analysis:** Gains primarily come from improvements in **Visual Correctness** and **Text Accuracy**, the two most critical components of K-Score.

### Performance on WISE Benchmark

**Table 2: Performance on the WISE benchmark (Overall Score).**
| Model | Overall Score |
| :--- | :--- |
| Qwen-Image | 0.62 |
| LongCat-Image | 0.65 |
| **Gen-Searcher-8B + Qwen-Image** | **0.77** |

Gen-Searcher improves Qwen-Image from **0.62 to 0.77** on WISE, a gain of 0.15, demonstrating strong generalization to other knowledge-based generation benchmarks.

### Ablation Study

**Table case: Ablation Study on KnowGen (K-Score with Qwen-Image).**
| Method | K-Score |
| :--- | :--- |
| Qwen-Image (Baseline) | 14.98 |
| + Manual Workflow (no training) | 22.91 |
| + Gen-Searcher-SFT only | 28.15 |
| + Gen-Searcher w.o. text reward ($\alpha=0$) | 29.59 |
| + Gen-Searcher w.o. image reward ($\alpha=1$) | 29.36 |
| **+ Gen-Searcher (Full)** | **31.52** |

*   **SFT is crucial:** Learning from trajectories is better than a manual workflow.
*   **RL provides further gains:** Beyond SFT initialization.
*   **Dual rewards are complementary:** Removing either reward leads to degradation, validating the design.

### Parameter Analysis
Performance remains strong when the balancing coefficient $\alpha$ is in the range **0.3 to."

## Theoretical and Practical Implications
*   **Advancing Agentic AI for Creative Tasks:** Demonstrates that agentic RL can be successfully applied to complex, creative tasks like image generation, moving beyond traditional QA or tool-use domains.
*   **Bridging the Knowledge Gap in Generative Models:** Provides a generalizable framework to augment any image generator with up-to-date, external knowledge without retraining the generator itself, addressing a fundamental limitation.
*   **Importance of Multimodal Search:** Highlights the necessity of retrieving both **textual evidence** and **visual references** for accurate generation in real-world scenarios, as text-only search is insufficient for fine-grained visual attributes.
*   **Robust RL Training Design:** The dual reward feedback mechanism offers a solution to the challenge of noisy rewards in end-to-end creative pipelines, making RL training more stable and effective.
*   **Foundation for Future Research:** The open-sourced datasets, benchmark, and model establish a foundation for developing more capable search agents for generation and other multimodal tasks.

## Conclusion
Gen-Searcher is the first trained multimodal deep search agent for knowledge-intensive image generation. By constructing novel datasets and a benchmark, and training via SFT and agentic RL with dual rewards, the model achieves substantial performance improvements and demonstrates strong transferability across image generators. This work opens a new direction for augmenting generative models with active, web-powered knowledge retrieval. Future work may explore scaling to larger models, extending to video generation, and improving the efficiency of the search-and-generation pipeline.

---

_Markdown view of https://picx.dev/p/6nnvNK, served by PicX — AI-generated visual whiteboard summaries of research papers._