Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
Jasper-Token-Compression-600Mhuggingface.co · Low evidenceJasper Token Compression 600M is a bilingual Chinese-English vector model that dynamically compresses text tokens by up to 10x while maintaining strong retrieval performance. It combines vector distillation and contrastive learning to improve efficiency, though retrieval performance still lags behind teacher models.
Parse Score