We present the Findings of the WMT 2026 Shared Task on Multitask LLMs with Limited Resources. This year, we extended the Shared Task to jointly model five diverse NLP tasks, Machine Translation (MT) for five language pairs (eight directions), Question Answering (QA), Spell Checking (SC), Grammar Checking (GC), and Mathematical Reasoning (MR), for two language tracks: Ukrainian and Sorbian (grouping Upper and Lower Sorbian). As the Shared Task is focused on low-resource conditions, the model is restricted to Qwen3.5-2B to remain below the 3B-parameter threshold.

In total, six teams participated across the two language tracks: two in the Ukrainian and four in the Sorbian track. No team participated in both language tracks. All systems have been uploaded to HuggingFace by the participants. Despite our strict restrictions on model and dataset availability, systems were diverse in the resources they used as well as their training and inference strategies. The submitted systems show that a single small model can handle all five tasks, with clear difficulty shown for the challenging MR only.

LLM-based translation remains challenging for low-resource languages, as LLMs often lack linguistic knowledge due to their limited representation in pre-training data. This is particularly problematic for morphologically rich languages where a large vocabulary can lead to further data sparsity.

Previous work has shown that providing external guidance on the source sentence in the form of language-specific information can help LLMs with tasks such as machine translation. In this study, we investigate zero-shot context augmentation for Upper Sorbian-to-German translation by enriching prompts with different types of contextual information: translation candidates in the form of bilingual dictionary entries, and morpho-syntactic analysis provided through morphological tagging. We first compare the effectiveness of these types of information in a zero-shot setting and then evaluate whether they generalize across additional experimental settings. We find that context-augmented prompting can improve Upper Sorbian-to-German translation. Contexts only containing translation candidates perform overall best, whereas morpho-syntactic analysis offers limited benefits, both on its own and in combination with lexical cues.

LLMs have shown impressive linguistic competencies in a wide range of NLP tasks, yet morphological generalization remains challenging. Models can benefit from in-context learning, where the choice of few-shot examples significantly influences the outcome. We probe LLMs’ morphological capabilities by assessing how two targeted in-context exemplar selection strategies, Iterative Confidence-Based Exemplar Selection (i-CES) and Similarity-Based Exemplar Selection (SES), influence the model’s ability for inflectional and derivational modeling across multiple languages. Our experiments with LLaMA 3.3-70B-instruct and Qwen2.5-32B-Instruct models indicate that SES is more effective for accessing latent linguistic knowledge, showing a substantial performance advantage over i-CES in both derivational and inflectional tasks. In contrast, i-CES offers no clear advantage over random sampling.

Byte Pair Encoding (BPE) is blind to morphology, a poor fit for Arabic, whose words bundle conjunctions, articles, and inflection into one token. The leading morphology-aware strategy for Arabic, MorphBPE, relies on Farasa, a supervised segmenter disallowed under the BabyLM 2026 rule against off-the-shelf linguistic tools. We study Arabic segmentation under the BabyLM GPT-2 training pipeline and propose Morfessor+MorphBPE, an unsupervised morphology-aware tokenizer.

We compare two dimensions: supervised vs. unsupervised morphological segmentation and the merging constraints used by BPE. All tokenizer variants cluster within a narrow accuracy band, with improvements rather on language understanding tasks than knowledge tasks, even though we do not observe consistent improvements over BPE. Furthermore, we find that holding the Farasa segmentation fixed, the MorphBPE constraint sharply lowers perplexity but barely changes downstream accuracy; and unsupervised Morfessor boundaries match supervised Farasa on aggregate while scoring higher on grammaticality (MultiBLiMP). Morfessor+MorphBPE thus offers a practical path to morphology-aware tokenization for Arabic.

We present the findings of the WMT 2025 Shared Task LLMs with Limited Resources for Slavic Languages. This shared task focuses on training LLMs using limited data and compute resources for three Slavic languages: Upper Sorbian (hsb), Lower Sorbian (dsb), and Ukrainian (uk), with the objective to develop and improve LLMs for these languages. We consider two tasks which are to be evaluated jointly: Machine Translation (MT) and Multiple-Choice Question Answering (QA). In total, three teams participated in this shared task, with submissions from all three teams for the Sorbian languages and one submission for Ukrainian. All submissions led to an improvement compared to the baseline Qwen2.5-3B model through varying fine-tuning strategies. We note, however, that training purely on MT degrades original QA capabilities. We also report further analyses on the submissions, including MT evaluation using advanced neural metrics for Ukrainian, as well as manual annotation and comparison to the current Sorbian machine translator.

This paper presents a study of the linguistic knowledge and generalization capabilities of Large Language Models (LLMs), focusing ontheir morphosyntactic competence. We design three diagnostic tasks: (i) labeling syntactic information at the sentence level - identifying subjects, objects, and indirect objects; (ii) derivational decomposition at the word level - identifying morpheme boundaries and labeling thedecomposed sequence; and (iii) in-depth study of morphological decomposition in German and Amharic. We evaluate prompting strategies in GPT-4o and LLaMA 3.3-70B to extract different types of linguistic structure for typologically diverse languages. Our results showthat GPT-4o consistently outperforms LLaMA in all tasks; however, both models exhibit limitations and show little evidence of abstract morphological rule learning. Importantly, we show strong evidence that the models fail to learn underlying morphological structures. Therefore,raising important doubts about their ability to generalize.

We empirically study the ability of a Large Language Model (gpt-3.5-turbo-instruct) to understand morphologically complex words. In our experiments, we looked at a variety of tasks to analyse German compounds with regard to compositional word formation and derivation, such as identifying the head noun of existing and novel compounds, identifying the shared verb stem between two words, or recognizing words constructed with inappropriately used derivation morphemes as invalid. Our results show that the language model is generally capable of solving most tasks, except for the task of identifying ill-formed word forms. While the model demonstrated a good overall understanding of complex words and their word-internal structure, the results also suggest that there is no formal knowledge of derivational rules, but rather an interpretation of the observed word parts to derive the meaning of a word.

The role of subword segmentation in relation to capturing morphological patterns in LLMs is currently not well explored. Ideally, one would train models like GPT using various segmentations and evaluate how well word meanings are captured. Since this is not computationally feasible, we group words according to their segmentation properties and compare how well a model can solve a linguistic task for these groups. We study two criteria: (i) adherence to morpheme boundaries and (ii) the segmentation consistency of the different inflected forms of a lemma. We select word forms with high and low values for these criteria and carry out experiments on GPT-4o’s ability to capture verbal inflection for 10 languages. Our results indicate that in particular the criterion of segmentation consistency can help to predict the model’s ability to recognize and generate the lemma from an inflected form, providing evidence that subword segmentation is relevant.

We study whether linguistic information in pre-trained multilingual language models can be accessed by human language: So far, there is no easy method to directly obtain linguistic information and gain insights into the linguistic principles encoded in such models. We use the technique of prompting and formulate linguistic tasks to test the LM’s access to explicit grammatical principles and study how effective this method is at providing access to linguistic features. Our experiments on German, Icelandic and Spanish show that some linguistic properties can in fact be accessed through prompting, whereas others are harder to capture.

We present the findings of the WMT2022Shared Tasks in Unsupervised MT and VeryLow Resource Supervised MT with experiments on the language pairs German to/fromUpper Sorbian, German to/from Lower Sorbian and Lower Sorbian to/from Upper Sorbian. Upper and Lower Sorbian are minoritylanguages spoken in the Eastern parts of Germany. There are active language communitiesworking on the preservation of the languageswho also made the data used in this Shared Taskavailable.In total, four teams participated on this SharedTask, with submissions from three teams for theunsupervised sub task, and submissions fromall four teams for the supervised sub task. Inthis overview paper, we present and discuss theresults.

This paper summarizes the results of our test suite evaluation with a main focus on morphology for the language pairs English to/from German. We look at the translation of morphologically complex words (DE–EN), and evaluatewhether English noun phrases are translated as compounds vs. phrases into German. Furthermore, we investigate the preservation of morphological features (gender in EN–DE pronoun translation and number in morpho-syntacticallycomplex structures for DE–EN). Our results indicate that systems are able to interpret linguistic structures to obtain relevant information, but also that translation becomes more challenging with increasing complexity, as seen, for example, when translating words with negation or non-concatenative properties, and for the morecomplex cases of the pronoun translation task.

This paper studies strategies to model word formation in NMT using rich linguistic information, namely a word segmentation approach that goes beyond splitting into substrings by considering fusional morphology. Our linguistically sound segmentation is combined with a method for target-side inflection to accommodate modeling word formation. The best system variants employ source-side morphological analysis and model complex target-side words, improving over a standard system.

This paper presents a simple method for German compound splitting that combines a basic frequency-based approach with a form-to-lemma mapping to approximate morphological operations. With the exception of a small set of hand-crafted rules for modeling transitional elements, this approach is resource-poor. In our evaluation, the simple splitter outperforms a splitter relying on rich morphological resources.

Many errors in phrase-based SMT can be attributed to problems on three linguistic levels: morphological complexity in the target language, structural differences and lexical choice. We explore combinations of linguistically motivated approaches to address these problems in English-to-German SMT and show that they are complementary to one another, but also that the popular verbal pre-ordering can cause problems on the morphological and lexical level. A discriminative classifier can overcome these problems, in particular when enriching standard lexical features with features geared towards verbal inflection.