Contents
Nowadays, pretty much all of us access the AI model that we trust on a daily basis. A tool like Google Gemini or ChatGPT acts as an all-purpose weapon that can do pretty much anything.
Here’s the catch: If we don’t use the paid Pro models, we have to live with two different hurdles. Either I end up with a less capable AI as the premium version is behind the paywall, or I can tap into the full intelligence, but I will have to live with very strict limits.
That’s exactly why I’m introducing you to lmarena.ai! You don’t have to sign up for a new account, you don’t pay a single penny, and you have a whole cornucopia of state-of-the-art models to choose from at any time. The perfect AI is always just one click away!
What is LMArena?
LMArena is a freely accessible online platform where you can try and compare different AI models. You can generate texts, conduct dialogs, create images, or request code — free of charge, and in many cases, even without registration. The focus is on easy access to different AI systems via a common interface.
Why LMArena is Suitable for Everyone
In everyday life, it’s often not about finding the best AI model, but about getting a result quickly. This is exactly what LMArena is perfect for. If you need a text draft, want to generate an image, or are looking for help with programming, you can perform it quickly here.
Instead of clicking through individual tools or testing several services in parallel, LMArena bundles these options in one place. This saves time and lowers the barrier to entry — especially for users who do not use AI regularly.
LMArena Modes at a Glance
LMArena offers three central usage types that are clearly differentiated from one another. You can find them at the top left of the window, with the default setting being battle mode.
Battle Mode
In battle mode, you enter a prompt, which is then executed by two nameless models. Only when the results are available and you can decide which of the two AIs has delivered the better result, will the names of the two models be displayed. You have no influence on which models compete against each other. In short, it will be a blind test.
However, it can happen this way: behind the initially nameless platforms, there is a brand new model that has not yet been released officially. The most prominent example of this is probably the image generation capabilities of Google Gemini, which you could already use under the code name Nano Banana before Google officially unveiled it.
Side-by-Side Mode
The side-by-side mode works in a similar manner but here, you select the two tools that are to compete against each other. The still unofficial platforms are missing here, but apart from that, you have many choices.
Direct Chat Mode
Finally, the direct chat mode works like a classic AI chatbot. Instead of a comparison, you work with a single model and can ask questions, refine answers, or develop tasks step by step. This is the suitable mode if you want to work specifically on a text, generate code, or conduct a longer dialogue.
Incidentally, in the above image, you can see not only the available modes on the right, but also the leaderboard on the left, i.e. the ranking list of available models. Below that, you can see in my history that I last used the Google AI once, and before that, I had Google compete against ChatGPT.
How to Use LMArena in Reality
The use of LMArena is very simple. After accessing the website, you can select a mode directly. The desired prompt is then entered in the input line as in any other AI chatbot — such as a task or an image description.
You can also click on the icons there, depending on whether you want to generate an image, video or code, or perform a web search. If you want, you can also upload images by tapping on the plus sign on the left. Depending on the mode, LMArena then provides either a single answer or several results for comparison.
Registration is not mandatory for many functions. This makes the platform suitable for quick use on the go, whether on your desktop or smartphone. The results can be used directly, copied, or used as a basis for further queries.
The Aquarium Example
In the following image, you can see why the quality of the AI can make a difference. I pitted the latest version of Claude against an older one. The selected task was as follows:
There are 10 fish in an aquarium. 2 of them drown, 3 swim away and 4 die. How many fish are left in the aquarium?
The reason this sentence works so well is that it reveals quite clearly what an AI is capable of. A bad AI model simply calculates blindly (10 – 2 – 3 – 4 = 1).
An average model comes up with 6 – the AI knows that fish do not swim away or swim out of closed containers, but counts the four dead fish even though they are still in the aquarium. The clever model provides the correct answer: 10!
PS: If you want to create images with Google Gemini, there are two more advantages: Firstly, the visible watermark disappears. In addition, Google, in certain countries like Germany is usually reluctant to generate images with celebrities. LMArena usually barrels its way through such restrictions!
A Real Game Changer
LMArena does not see itself as a replacement for specialized AI services, but as low-threshold access for the hoi polloi. If you are thinking about which tool is right for you in the long term or simply need occasional AI support, you will find a flexible solution here. Instead of committing yourself early on, you can try out, compare, and decide what works best for your own use case.
It is precisely this openness that makes LMArena a practical tool for everyday AI: free, versatile and with no obligations. If you ask me, that’s a real game-changer.
