HexaBee-4B hits 73.75 chrF++ across six languages, beating Gemma 3 12B at a third the size
Iman Barati · hf · 2026-10-14
Researchers from Iran University of Science and Technology present HexaBee-4B, a six-language translator (Arabic, Chinese, English, Hebrew, Persian, Spanish) adapted from Gemma 3 4B IT in three stages:
- Wikipedia continued pretraining.
- Joint multilingual answering and answer-text translation: 19.46M answering records, including 5.76M cross-language QA pairs and 10.70M answer-text translations.
- Bidirectional translation specialization on 658,274 synthetic examples.
Across 36,572 test examples spanning all 30 directions, it achieves direction-macro chrF++ of 73.75 vs 57.11 for Gemma 3 4B IT and 66.83 for Gemma 3 12B IT, with BLEU, TER, COMET and model-judged subsets confirming the ranking.
More from Research
- Laude Institute launches FrontierCS 2: agents must beat deleted paper contributions blind — lateinteraction · 2026-10-23
- RFD4-Proteina's 'underrated' docs detail partial diffusion for protein redesign — AllThingsApx · 2026-10-23
- MPI-Sec PhD Thesis on Vulnerability Detection Benchmarks Includes Pre-LLM Classic "Top Score on the Wrong Exam" — mboehme_ · 2026-10-23
- LiveNerf's 29-day data finds no nerfing of Claude Opus 5.5 — TheOnlyVibemaster · 2026-10-23
- New paper makes 9 falsifiable predictions toward a major AI-enabled terror attack by 2027-2028 — StephenLCasper · 2026-10-23
- HumanoidPose recovers humanoid robots' full-body joint states from RGB video alone — cholz · 2026-10-23