We present the findings of the WMT 2026, which evaluated bidirectional translation between Arabic and English, Hindi, Bangla, Indonesian, and Urdu. We release AraAsian1.0, a news-domain parallel corpus with 20.1K-21.0K training pairs per language pair, and evaluate submissions on ten translation directions using BLEU, ChrF2, TER, COMET, and BERT-based metrics. Of 23 registered teams, 13 submitted systems, producing 70 primary and 65 contrastive runs. Fine-tuned multilingual MT models remained the strongest overall: MADLAD-based systems won four primary directions, while target-aware pivoting, joint multilingual adaptation, and MBR-based ensembling led the remaining tracks. The best contrastive run exceeded the best primary run in only three of ten directions, and each gain was below 0.2 BLEU, indicating that additional data, prompting, or multi-pass refinement did not provide consistent improvements. Bangla had the lowest winning BLEU in both directions, despite comparable training-set size. BLEU and COMET selected the same winner in nine of ten primary results but only six of ten contrastive results, with the largest disagreements occurring for LLM-based and reranked outputs. These results emphasize domain-matched adaptation, target-aware transfer, and multi-metric reporting in low-resource Arabic-Asian translation. We made the AraAsian1.0 dataset available for the community.

In recent times, internet users are frequently exposed to Hate Speech on social media platforms that have long-lasting negative impacts on their mental wellbeing and also radicalizes the society into an environment of fear and distrust. Many methods have been developed to detect and stop propagation of Hate Speech. However, there is a limitation of annotated data available for Hate Speech in Bengali language. In this work, we have used a pretrained BanglaBERT model on an extended train dataset synthesized via data augmentation techniques. Our team Bahash-AI has achieved 20th, 20th and 17th position of the 3 subtasks out of total 37, 24 and 21 total number of teams who participated in the subtasks 1A, 1B and 1C respectively for Bangla Multi-task Hatespeech Identification Shared Task at BLP Workshop with F1 scores of 0.7028, 0.6954, 0.6969 respectively.

The neural machine translation approach has gained popularity in machine translation because of its context analysing ability and its handling of long-term dependency issues. We have participated in the WMT21 shared task of similar language translation on a Tamil-Telugu pair with the team name: CNLP-NITS. In this task, we utilized monolingual data via pre-train word embeddings in transformer model based neural machine translation to tackle the limitation of parallel corpus. Our model has achieved a bilingual evaluation understudy (BLEU) score of 4.05, rank-based intuitive bilingual evaluation score (RIBES) score of 24.80 and translation edit rate (TER) score of 97.24 for both Tamil-to-Telugu and Telugu-to-Tamil translations respectively.