Because it's Google they can't build products and they only care about benchmarking.
The product they released so far are all half assed experiments.
Gemini 3 Pro is now being beaten by open source models because they can't fix or don't want to fix the problems with the Gemini models being completely useless.
The same for Microsoft.
Microsoft had GitHub Copilot, and Microsoft Copilot and both of them are useless to Claude Code and Claude Cowork.
You can have all the money in the world, but nothing is stopping you from building useless garbage.
Claude is clearly the most superior product right now.
Gemini is absurdly expensive for low quality (3000 USD of tokens are not even worth what you get @ Anthropic for 200 USD).
The same can be said about Claude (no or tiny Opus on Pro) vs GPT-5.2 high (5.3-codex if you like terminal bench hacking).
You mean GPT-5.3-Codex is a much better value than Claude Opus for programming ? If yes then I'm very interested as I am using Claude there
I guess it depends on your spending. GPT-5.2 and -5.3-Codex are certainly much cheaper: you get much more from the same $20 sub. When I was using Claude as primary I would daily hit limits and have to wait vs on GPT with more usage it only happened to me once in a few months when I was vibecoding non-stop for a week or two to port my personal Windows tools to Linux with multiple other projects being worked on in parallel.
Anecdotally GPT was also smarter than Claude which prompted my move from Claude in the first place: Gemini and Claude back in October failed to get their own harness PID.
Outside of anecdata I rely on https://artificialanalysis.ai/models/capabilities/coding for now.
I really like github copilot.
I also tried open code cli and desktop, but how well copilot is integrated into the ide is a plus for me.
What makes them "useless garbage"?