Ant Group opensources Ling-3.0 Base models with training checkpoints
智东西 · wechat · 2026-08-20
Ant Group's Bailing team opensourced Ling-3.0-tiny-base (7.9B) and Ling-3.0-flash-base (124B), releasing 6 checkpoints covering pre-training, mid-training, and WSM merging stages. The Tiny model (1.3B active) targets code and RL research, while Flash (5.1B active) focuses on code, long context, and reasoning. They use a WSM learning rate scheme for better continual pre-training.
Related event: Ant Group Open-Sources Ling-3.0 Base Models With Full Training Checkpoints(4 posts)→
More from Models
- Sanity Check: Claude $100 Seat Equals ~$2500 in API Credits for Heavy Users — weed_cutter · 2026-08-21
- Claude accurately predicted Qwen 3.8 27B performance benchmarks — OneMoreName1 · 2026-08-21
- GLM 5.3 Scores 47.1% on SlopCodeBench, Ties with Fable 5 — corruptbytes · 2026-08-21
- Post-training causes LLMs to produce novel but impractical language — TuhinChakr · 2026-08-20
- User says Grok 4.6 now handles 100% of coding work previously done with Codex — CedricMakes · 2026-08-20
- Grok 4.6 leads in legal/GDP benchmarks, lags in coding — ChrisGPT · 2026-08-20