Aug 11 — Bailing said it has open-sourced Ling-3.0-tiny, a native mixed-precision inference model with 7.9 billion total parameters and only 1.3 billion activated parameters at inference. The release includes BF16, FP8 and INT4 builds.

2026-08-11

Aug 11 — Bailing said it has open-sourced Ling-3.0-tiny, a native mixed-precision inference model with 7.9 billion total parameters and only 1.3 billion activated parameters at inference. The release includes BF16, FP8 and INT4 builds.