A derivative model cuts reasoning overhead by roughly 40 percent while holding accuracy steady, and it ships only…
The new jina-ocr-v1 turns PDFs, scans and tables into clean Markdown on a single low-cost GPU.
SPARSEUP packs a 149M-parameter encoder under Apache 2.0 weights and reports 56.4 average nDCG@10 on BEIR-13.
Inkling, a 975-billion-parameter open-weight model, lets organizations customize AI without relying on big labs.