Ling-2.5-1T: a trillion parameters, 63B active, one million token context
- Source
- x.com
- Date
Meet Ling-2.5-1T! 📊1T params (63B active), 29T pre-training corpus & 1M context #LLMs #OpenSource As the latest flagship instant model in the @AntLingAGI family, Ling-2.5-1T delivers comprehensive upgrades across model architecture, token efficiency, and preference alignment. 🤗 Hugging Face: https://t.co/MEGiB6X6Ke 📷 ModelScope: https://t.co/VRWgfKxW7c
Context
Running a model with a million- normally means either accepting a huge active-parameter cost at time or using a much smaller model that can't hold as much context. Ant Open Source's Ling-2.5-1T, part of the Ant Ling family, activates 63 billion of its trillion total parameters per token, trained on a 29-trillion-token corpus, while supporting a 1 million token context window. Ant describes this release as carrying upgrades across model architecture, token efficiency and preference relative to earlier Ling models, with weights published on Hugging Face and ModelScope.
- token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
- context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
- inference — Running a trained model to get answers — the phase where AI is actually used, as opposed to trained.
- alignment — The work of making AI systems actually pursue what their builders and users intend, rather than something subtly or dangerously different.
Checking sign-in…
Loading comments…