I first met AI when OpenAI released ChatGPT in November 2022. Other launches followed: in February 2023 Microsoft put a chat into Bing, in March Anthropic released Claude, and Google opened up Bard, which was later renamed Gemini.
Since then AI has become an inseparable part of my everyday work.
By now there are a great many companies building AI, and even more models out there. You can see what the market has to offer on Artificial Analysis.
Artificial Analysis
Artificial Analysis is an analytics platform that tests and compares AI models and the services that run them. The project positions itself as an independent tester. On the site you can go through test results and compare models by answer quality, speed and cost. It is a handy reference point for making sense of the sheer number of models and understanding how they differ from one another.
In the end, my favourites in 2026 are ChatGPT from OpenAI and Claude from Anthropic. As for specific models, at the time of writing those are GPT-6 Astra and Claude Fable 5.1 — they cover 99% of what I need.
I use the models for all sorts of things: development, working with text, images, video and documents, everyday questions and so on. I also use cheap Chinese models for API integrations, but that is a whole other story.
