Gemini 3, and a three-way race with no clear winner
Google released its new generation yesterday, a week after OpenAI’s update. For the first time the honest answer to “which is best” is “it depends what for”.
Google released Gemini 3 on 18 November 2025, six days after GPT-5.1 and a week before Anthropic’s next release. Three major providers are now within touching distance of each other on most measures.
For a Swiss SME this is more significant than any one of the three releases, and for a reason that has nothing to do with capability.
What a genuine three-way race produces
Price pressure, which you have already seen: prices fell substantially in June and again through the autumn. Feature parity, which means the thing you liked about one provider tends to appear at the others within two quarters.
And, most usefully, negotiating room. A supplier who knows you have two equivalent alternatives behaves differently from one who does not.
The question that has quietly become unanswerable
« Which model is best? » No longer has a useful answer, and anyone who gives you one confidently is describing either a benchmark you do not care about or a preference.
The answerable version is narrower: which is best for the twenty examples from last Friday’s post, in German and Italian, under a contract you can accept. That question has an answer and it is yours specifically.
What we would actually do in the next four weeks
Nothing, for most companies. Three launches in a fortnight is a reason to wait for the dust rather than to act during it.
If you are mid-decision, note that the market has just become more competitive, which means an offer made in January will be better than one made now. That is a legitimate reason to let a conversation run into the new year.
Friday’s post is about how to sit through a month like this without making a decision you will regret in February.