AI Test Environments

AI test environments

Which AI language model is best suited to your use case? How can existing AI models be adapted for use in your own company? WestAI offers two digital test environments. The LLM Lab enables the secure and data protection-compliant testing of various large language models, both open and closed source. The AI.Lab offers a flexible environment in which AI solutions can be individually adapted and developed to meet specific company needs. Try it out for yourself!

The right AI language model for every scenario

ChatGPT, Gemini oder Teuken? Welches KI-Sprachmodell ist das Richtige im Unternehmenskontext? Das LLM Lab von WestAI hilft bei der Auswahl. Als leistungsstarke und zugleich benutzerfreundliche Plattform ermöglicht es das LLM Lab, große Sprachmodelle (engl. Large Language Models, LLMs) miteinander zu vergleichen und gezielt auszuprobieren. Es bietet Unternehmen den Zugang zu einer breiten Auswahl an Open-Source- und Closed-Source-Modellen – zentral gebündelt in einer sicheren, datenschutzkonformen Umgebung. Effizient entscheiden, gezielt einsetzen Um das passende Modell für den Unternehmenseinsatz zu finden, müssen verschiedene Faktoren oder Benchmarks berücksichtigt werden. Entscheidend ist z. B., wie ein Modell im konkreten Anwendungsfall performt – mit echten Daten, Fachsprache und Zielgruppen. Ebenso spielt die Art der Promptgestaltung (Prompt Engineering) eine zentrale Rolle, da sie die Qualität und Relevanz der Modellausgaben maßgeblich beeinflusst.

Try it out now

This is where WestAI’s LLM Lab comes in:

Compare different models under real conditions,
identify the right model for your specific requirements,
Optimize prompts specifically for maximum performance.

With functions such as direct model comparison and a clearly structured interface, the LLM Lab enables well-founded decisions – whether for automation, customer communication or internal processes. The most important functions at a glance:

Comparison of AI language models

Direct comparison of different large language models, both open source and proprietary, in terms of their response quality to specific inputs or prompts.

Seamless integration into customer applications

The developed and optimized prompts can be seamlessly integrated into customer applications or projects (import and export of generated prompt templates.

Data security

The use of open source models in a protected environment prevents important data from being passed on to external API services.

Model Cards

Comparison of the latest open source models in terms of their key figures, such as model size, supported languages or available licenses.

Prompt Engineering

Users can experiment with the formulation of prompts to optimize the performance and responses of the models.

Intuitive operation

Our user interface is particularly user-friendly and can be used without in-depth technical (AI) knowledge.

Testing use cases and new ideas

Test and compare models for customer-specific use cases, data, languages or new ideas.

Figure 1: Compare Chat LLM’s

Figure 2: Screen recording of the LLM Lab

Which AI language models are available? The current version of the LLM Lab provides access to variants of the multilingual open source Teuken model family developed by the OpenGPT-X project according to European needs, as well as other publicly available open source models. Users can also access commercially available proprietary models, such as GPT models from OpenAI, if they provide their own API key.

To the LLM Lab

The AI.Lab powered by KI.NRW

Try out your AI on your own

In the AI.Lab, users from companies and research institutions can quickly and easily try out their own multimodal AI models or those developed with us in the public AI playground. Alternatively, the use cases can also be run locally in the protected IT environment of your own institution.

WestAI particularly improves the interaction of the AI Lab with supercomputing environments in order to simplify use cases such as fine-tuning or “training from scratch”. The modular principle allows the AI modules to be reused efficiently. Large language models alone are often not enough to achieve the desired quality of answers. They must therefore be combined with other AI technologies, such as knowledge graphs (Figure 1) or domain-specific collectors (Figure 2).

Our experts are also available to assist here on request. As an open source platform, the AI.Lab is constantly evolving. Users are therefore welcome to get involved and make suggestions for improvements and feature requests, for example.

Figure 1: Language model and knowledge graph

Figure 2: Speech model and collector

The AI.Lab is operated by KI.NRW and our partner Fraunhofer IAIS. It is part of the European AI on Demand (AIoD) initiative and is supported by the European Commission and the Ministry of Economic Affairs, Industry, Climate Action and Energy of the State of North Rhine-Westphalia.

To the AI.Lab