Tag: llm
-
Tencent Just Open-Sourced a 770-Billion Model
Tencent just gave away one of the biggest AI models anyone has built, for free. The company open-sourced Hy4 preview, a model built on 770 billion parameters. It only lights up 49 billion of those for any given task, which is the trick that keeps it fast and affordable instead of impossibly expensive to run.…
-
xAI Just Shipped Grok 4.6, and 4.7 Is Weeks Away
xAI shipped Grok 4.6 today, and if you are keeping track, that is the second major Grok release in about six weeks. Grok 4.5 had barely finished rolling out before Elon Musk was already promising the next one, and here it is, right on schedule. Grok 4.6 runs on the same 1.5 trillion parameter foundation…
-
OpenAI’s Three-Tier GPT-5.6 is Now Live
OpenAI just shipped GPT-5.6 to everyone, and they did it with three different models instead of one. That is the interesting part. Sol is the frontier model. It costs five times more than Luna per token, and it is built for the work that needs raw reasoning power over long horizons. Agentic tasks. Complex problem-solving.…
-
Grok 4.5 is here and it’s aiming straight at Claude
SpaceXAI just launched Grok 4.5 to the public on July 9, and the first thing you notice is not the benchmarks. It is that SpaceXAI is playing a completely different game than it was three months ago. Grok 4.5 is an Opus-class model trained on a 1.5 trillion parameter foundation built with Cursor coding data.…