Announcing Grok 4.5, our first model trained specifically for coding and agents. It was trained with Cursor and offers frontier intelligence at leading speeds and cost efficiency.
https://t.co/i8HpU7w64k
@MartinBergo@skcd42 Hey feel free to DM me!
https://t.co/exjo6f4SKC
Have some docs on custom models here. You just need to define the model config in your ~/.grok/config.toml
@karrisaarinen@sean_lynch I did something similar with Amazons internal code review tool. It ended up with a few thousand engs using it internally
You can get pretty far with tampermonkey
@jeffreyhuber@darrenjr when models can do better at long horizon experimentation like
“Evaluate the performance difference between a FastAPI server and Axum for our new XYZ service”
the need for stronger isolation between concurrent runs will be much more important