News

It's easy to rely on a single number from a leaderboard when choosing a model for your use case. But that number can hide far more than it reveals. It's important to also look at how informative the benchmarks used actually are, and what capabilities they evaluate.

One of the main goals of SLAIH is to make it easier to discover high-quality resources for Slovak language AI. In this edition, we'd like to highlight a few recent additions that we believe deserve your attention. Whether you're building an LLM application, training an ASR system, or experimenting with multilingual models, these resources are well worth exploring.

If you want to know how well speech-to-text models work for English, you have your answer in a minute. For Slovak, until recently you had no answer at all. Multilingual evaluations either skip Slovak entirely or include it as a fraction of a percent of the test data, far too little to say anything reliable.

We are releasing the first dedicated automatic speech recognition benchmark for Slovak on SLAIH.

SkMTEB — the first comprehensive benchmark for text embeddings in Slovak — has been accepted at ACL 2026, one of the most prestigious international conferences in natural language processing. The benchmark covers 31 datasets across 7 task types and establishes a standard comparable to MTEB for English. The work is led by a team from the Kempelen Institute of Intelligent Technologies (KInIT), the Technical University of Košice (TUKE), and Comenius University Bratislava (UK), together with other partners from the Slovak AI ecosystem.

We are launching Slovak Language AI Hub - the central place for datasets, models, and benchmarks for the Slovak language. Our goal is to improve the availability of resources and support the development of NLP in Slovakia.