Rémy @xowap.dev · Apr 28 So according to the Terminal-Bench Hard benchmark on Artificial Analysis, Gemma 4 26B A4B could roughly sit at the same level as Gemini 3 Flash, which I've found largely sufficient for most of my coding needs. Can't wait to see if it runs on my GPU.
0 likes 1 replies
?
Replies
Also Gemma 4 E4B beats Ministral 14B, which was already a pretty useful model for many tasks. Although I need to remember that I don't believe in benchmarks. But still. It means by ~2030 you will easily be able to run local coding models on your local machine. Don't buy AI datacenters just yet 🤓