Published on April 3, 2026
EXAONE 4.5: LG Releases Its First Open Multimodal Language Model
LG Research has introduced EXAONE 4.5, an open multimodal language model capable of simultaneously analyzing text and images.
Products 3 – 5 minutes min readMajor tech companies are increasingly releasing their language models as open source. Following Google, Meta, and several other players, LG has now done the same. The South Korean tech giant is primarily known to most people for its TVs and home appliances. The company's research division, LG AI Research, has released the EXAONE 4.5 model – LG's first creation to be made publicly available.
What is EXAONE 4.5?
Simply put, it's a language model that can work with both text and images simultaneously. In the industry, this class of models is called multimodal – or, as seen in some publications, a Vision Language Model. The model can interpret an image and reason about its content, combining visual context with a text prompt.
Before the release of EXAONE 4.5, the company had already been developing the EXAONE series, but those versions remained proprietary – used internally at LG and in partner projects. This is the first time LG has opened a model to the general public.
Why This Matters – Even If You're Not a Programmer
When a major company releases a model as open source, it means that developers, researchers, and simply curious users can study it, run it, integrate it into their own products, or test it on their own tasks – without paid subscriptions or closed APIs.
For the industry, this sends a clear signal: LG is not just investing in AI but is also ready to compete in the open model space. This benefits everyone – more open models mean more options, greater independence from a single provider, and more opportunities for experimentation.
What the Model Can Do
EXAONE 4.5 is positioned as a model with advanced reasoning capabilities. This isn't just about generating template-based answers – the model can construct logical chains, analyze data, and handle complex queries that require multiple steps of thought.
In terms of practical scenarios, the model can analyze documents with charts and diagrams, answer questions about the content of images, and assist with tasks where text and visual content are intertwined. This is particularly relevant for working with reports, presentations, technical drawings, or medical images.
It's important to note that the model was developed for real-world application in business and professional environments – not just as a demonstration of a research lab's capabilities.
LG in the AI Race – An Unexpected but Logical Player
At first glance, it might seem surprising that a manufacturer of refrigerators and OLED screens is releasing language models. However, LG has long had its own research division dedicated to artificial intelligence – LG AI Research. It works on models that find applications in various sectors of the company's business, from smart appliances to industrial solutions.
The EXAONE series has been in development for several years, and EXAONE 4.5 represents a kind of maturity point: the model is strong enough to move beyond the corporate perimeter and stand alongside the products that form the foundation of today's open AI landscape.
What Remains to Be Seen
It is still difficult to say how competitive EXAONE 4.5 will be in practice compared to already established open models. Much will depend on how actively the community begins to use it and which specific tasks it will solve better than the alternatives.
A separate issue is the terms of use. An “open” model and a “completely free” one are not always the same thing. Licensing details determine where and how the model can be used: in commercial products, for research, or only in non-commercial projects. It's worth clarifying this before building anything serious on it.
Nevertheless, the publication itself is a significant step. LG is entering the open language model space, and this expands the options for those looking for alternatives to products from the usual American and Chinese providers.
You May Also Like
Events are only part of the bigger picture. These materials help you see more broadly: the context, the consequences, and the ideas behind the news.
Alibaba introduces Qwen3.6-Plus, an enterprise AI model capable of independently developing code and analyzing visual content in real-world scenarios.
The Bangalore-based developer has released a multimodal model that understands speech, text, and images, supports India's major languages, and is capable of operating offline.
AI: Events
Technical context • Infrastructure
Google has released a new family of Gemma models, optimized in collaboration with NVIDIA for local execution – from compact edge devices to powerful workstations.
How This Text Was Created
This material is not a direct retelling of the original publication. First, the news item itself was selected as an event important for understanding AI development. Then a processing framework was set: what needs clarification, what context to add, and where to place emphasis. This allowed us to turn a single announcement or update into a coherent and meaningful analysis.
We openly show which models were used at different stages of processing. Each performed its own role — analyzing the source, rewriting, fact-checking, and visual interpretation. This approach maintains transparency and clearly demonstrates how technologies participated in creating the material.
The neural network studies the original material and generates a coherent text
Correction of errors, inaccuracies, and ambiguous phrasing
Generating a textual prompt for the visual model
Generating an image based on the prepared prompt
Want to dive deeper into the world
of neuro-creativity?
Be the first to learn about new books, articles, and AI experiments
on our Telegram channel!
