Open Source Model Open Source

Granite 4.0 3B Vision

Granite 4.0 3B Vision is an IBM vision-language model designed for enterprise-grade document data extraction, focusing on charts, tables, and key-value pairs.

5/10 MegaOne Score

Visit Granite 4.0 3B Vision →

Price Free / Open Source

Category Open Source Model

Model Open Source

Company IBM

Team Size 268,000

About Granite 4.0 3B Vision

Granite 4.0 3B Vision is a vision-language model (VLM) from IBM Research, released in March/April 2026, specifically engineered for enterprise-grade document data extraction. It excels at complex tasks such as converting charts into structured formats (CSV, Summary, Code), accurately extracting tables with intricate layouts into JSON, HTML, or OTSL, and performing semantic Key-Value Pair (KVP) extraction across diverse document types. The model is delivered as a 0.5B parameter LoRA adapter designed to run on top of the Granite 4.0 Micro (3.5B) language backbone, allowing for efficient dual-mode deployment for both multimodal and text-only workloads.

Why We Scored It 5/10

5/10

Granite 4.0 3B Vision serves a specific niche in enterprise document AI with Apache 2.0 licensing, but operates in a highly competitive space dominated by more capable models like GPT-4o and Gemini. While IBM's enterprise focus and open-source approach provide some value, the model's specialized nature and limited adoption signals limit its broader impact.

Features & Platform

Free Tier

Granite 4.0 3B Vision Coverage