Compact Multimodal Intelligence for Enterprise Documents

• Table Extraction: Accurately parsing complex table structures (e.g., multi-row, multi-column, etc.) from document images
• Chart Understanding: Converting charts and figures into structured machine-readable formats, summaries, or executable code
• Semantic Key-Value Pair (KVP) Extraction: Identifying and grounding semantically meaningful key-value field pairs across diverse document layouts

The model ships as a LoRA adapter on top of Granite 4.0 Micro, our dense language model, keeping vision and language modular for text-only fallbacks and seamless integration into mixed pipelines. It continues to support vision-language tasks comparable to producing detailed natural-language descriptions from images (e.g., “Describe this image intimately”). The model will be used standalone or in tandem with Docling to boost document processing pipelines with deep visual understanding capabilities.

How Granite 4.0 3B Vision Was Built

Granite 4.0 3B Vision’s performance is the results of three key investments: A purpose-built chart understanding dataset constructed via a novel code-guided data augmentation approach, a novel variant of the DeepStack architecture that allows high-detail visual feature injection, and a modular design that keeps the model practical for enterprise deployment.

ChartNet: Teaching Models to Truly Understand Charts

Charts present a challenge for vision-language models (VLMs) because understanding them requires jointly reasoning over visual patterns, numerical data, and natural language, a mixture most VLMs cannot handle well, especially when spatial precision matters—comparable to reading exact values off a line chart. To shut this gap, we’ve developed ChartNet: a million-scale multimodal dataset purpose-built for chart interpretation and reasoning, described intimately in our upcoming CVPR 2026 paper.

ChartNet uses a code-guided synthesis pipeline to generate 1.7 million diverse chart samples spanning 24 chart types and 6 plotting libraries [see Figure 1]. What makes it so distinctive is that every sample consists of 5 aligned components—plotting code, rendered image, data table, natural language summary, and QA pairs—providing models a deeply cross-modal view of what a chart means, not only what it looks like. The dataset also includes human-annotated and real-world subsets, filtered for visual fidelity, semantic accuracy, and variety.

The result’s a training resource that moves VLMs from merely describing charts to genuinely understanding the structured information they encode—with consistent gains across model sizes, architectures, and tasks.

Figure 1: ChartNet’s synthetic data generation pipeline.

DeepStack: Smarter Visual Feature Injection

Most VLMs inject visual information into their language model at a single point, which forces the model to handle each high-level semantics and fine-grained spatial detail concurrently. Granite 4.0 3B Vision takes a unique approach with DeepStack Injection: abstract visual features are routed into earlier layers for semantic understanding, while high-resolution spatial features are fed into later layers to preserve detail. The result’s a model that understands each what’s in a document and where—which is critical for tasks like table extraction, chart understanding, and KVP parsing where layout matters as much as content. For a full technical breakdown, see the Model Architecture section of the model card.

Modularity: One Model, Two Modes

Granite 4.0 3B Vision is packaged as a LoRA adapter on top of Granite 4.0 Micro, quite than as a standalone model. In practice, this implies the identical deployment can serve each multimodal and text-only workloads, mechanically falling back to the bottom model when vision isn’t required. This keeps enterprise integration straightforward without sacrificing performance.

How It Performs

Charts: Evaluated on the human-verified ChartNet benchmark using LLM-as-a-judge, Granite 4.0 3B Vision achieves the very best Chart2Summary (86.4%) rating amongst all evaluated models, including significantly larger ones [see Figure 2]. It also ranks second on Chart2CSV (62.1%), behind only Qwen3.5-9B (63.4%), a model greater than double its size.

Figure 2: Granite 4.0 3B Vision performance on chart2csv and chart2summary, compared against peer vision-language models using LLM-as-a-judge.

Tables: We evaluate table extraction in two settings: cropped tables (isolated regions) and full-page documents (tables embedded in complex layouts) [see Figure 3]. The benchmark suite includes TableVQA-extract (cropped table images), OmniDocBench-tables (full-page documents), and PubTables-v2 (each cropped and full-page settings). Models are tasked with extracting tables in HTML format and scored using TEDS, a metric that captures each structural and content accuracy. Granite 4.0 3B Vision achieves the strongest performance across benchmarks, leading on PubTablesV2 on each cropped (92.1) and full-page (79.3), OmniDocBench (64.0), and TableVQA (88.1) scores amongst all evaluated models.

Figure 3: Granite 4.0 3B Vision’s table extraction performance across cropped and full-page benchmarks (TableVQA-extract, PubTables-v2, OmniDocBench-tables), measured by TEDS.

Semantic KVP: VAREX is a benchmark specifically designed to discriminate between small extraction models, comprising 1,777 U.S. government forms spanning easy flat layouts to complex nested and tabular structures. Models are evaluated using exact match (EM), a strict metric that requires the model’s extracted key-value pairs to match the bottom truth. Granite 4.0 3B Vision achieves 85.5% EM accuracy zero-shot.

How one can Use It

Granite 4.0 3B Vision can operate either as a stand‑alone visual information extraction engine or as a part of a totally automated document‑processing pipeline with Docling. The model is designed to support scalable, accurate extraction across diverse document types and visual formats.

1. Stand‑Alone Image Understanding
Granite 4.0 3B Vision can run directly on individual images, making this selection useful for applications with existing workflows that need targeted visual extraction without modifying upstream systems. This offers easy integration into existing automation workflows and is suitable for lightweight, task‑specific tools (e.g., form parsers, chart analyzers, etc.).

2. Integrated Document Understanding Pipeline With Docling
Granite 4.0 3B Vision can be integrated seamlessly with Docling to support complete end‑to‑end document understanding. This mode can offer:

• Large‑scale processing of multi‑page PDFs
• Automated detection, segmentation, and cropping of figures, tables, and other visual elements with Docling and redirection of unpolluted crops to Granite Vision model for fine-grained extraction
• Efficient workflow with lower overall computational costs and faster throughput
• Higher accuracy, more reliable extraction, and significantly improved efficiency across large document collections

Example Use Cases

• Form Processing: Extract structured fields from invoices, forms, and receipts using KVP capabilities or generate natural‑language descriptions of figures using image2text feature (e.g., “Describe this image intimately”).
• Financial Report Evaluation: Use Docling to parse reports, detect figures, and crop visual elements. Process charts using Granite Vision’s chart2csv, chart2code, and tables using tables_json capabilities to convert them into structured, machine‑readable data enabling actionable insights.
• Research Document Intelligence: Utilize Docling to handle OCR and layout parsing across dense academic PDFs, and pass extracted figures to chart2summary and table crops to tables_html to make visual content discoverable alongside free-form text in a single pipeline.

Try It Today

Granite 4.0 3B Vision is obtainable now on HuggingFace, released under the Apache 2.0 license. Full technical details, training methodology, and benchmark results can be found within the model card. We’d love to listen to what you construct with it—share your feedback within the community tab.