They do not appear to care about the frontier. They're focusing on protecting their ad business, so their models need to be lightweight and efficient, to provide fast queries for people who are on Google, and for local inference on Android and Apple devices. Also specialised small models for things like OCR, translation, facial recognition, voice, design, etc. Their TPUs are also quite impressive.

It's a reasonable niche they're carving out, but I think too conservative. Very strange from Google. The other labs are racing to RSI and Google appears to have given up the race.

They're not until they are. No doubt they are working on something and have better resourcing than the AI labs

Grok has been at the frontier since 4.6

Gemini 4 Argon release a few days ago was frontier a few days ago. Things move fast.

I think some of the benchmarks that Meta didn't pick show Gemini 4 Argon ahead but really the benchmarks are not that great any more and getting increasingly disconnected from reality as people search for "headroom" in the evals to avoid saturation I think. Opus 5.5 is "top" for example but I really dislike the way it talks to me (it feels quite condescending, holier-than-thou, uptight etc) and prefer Gemini 3.8 flash as a daily driver just because it's so much more pleasant (and faster!) to work with and it's code is good-enough for my needs, despite being weaker in the benchmarks.