If you only look at the leaderboard, 31B is definitely the most eye-catching.
But when you actually get the machine out, it’s still that un-upgraded RTX 3060 12GB, and your judgment will change immediately. How should I put it? For local deployment, in the end, it’s not about who looks the fanciest, but who seems like the one you can live with long-term. For me, what is truly worth running first this time isn’t 31B, but 26B A4B.
On the day of the initial release, what I originally wanted to do was simple: find an upgraded version corresponding to Gemma 3 and download it to run.
However, after looking around, I was a bit stunned. The familiar naming convention of 4B / 12B / 27B is gone; instead, we have E4B, 26B A4B, and 31B. How should I put it? This time, what Google truly changed wasn’t just the model sizes, but even “how you should understand this batch of models.”
“Previously, it was mentioned that when logging into Gemini Cli, you needed to configure the Google Cloud Project ID. This is already wrong – if it’s a personal account, there shouldn’t be this restriction. The fact that this restriction exists indicates that you’ve started triggering Google’s security system and are being identified as not being a personal account.
It’s frustrating; after using it for half a month and getting used to it, now I have to return to the embrace of cc + domestic models.”
It’s been almost half a month, and MiniMax and GLM4.6 in China have both had paid trials – the difference is still noticeable. The cc tool is quite useful. Last night I was messing around with optimizing the front-end interface, you know… as a writer, I basically don’t understand front-end development; Vibe Coding will be next, and I’m starting to get into the front-end technology stack. Domestic large models haven’t been figured out; I tried the newly released Gemini3, and it was done in five minutes. Switching to the site’s archived pages, you can see “the bookshelf.”