China’s record open‑weight model is closing the frontier AI gap
What Changed and Why It Matters Chinese labs keep shipping very large open‑weight models that are cheap to run and easy to customize. The lag to frontier performance has shrunk…
A 975B open‑weight model just dropped. Here’s why it matters.
A new 975B‑parameter, open‑weight model called Inkling just launched. It’s downloadable, self‑hostable, and fine‑tunable. This isn’t just a model release. It’s another strong signal that open‑weight LLMs are becoming the…
Mistral bets on small models: stronger AI with less compute
What Changed and Why It Matters Mistral launched the Mistral 3 family with a clear stance: go smaller, run faster, and ship everywhere. The company released open‑weight frontier and compact…