Opus 4.6 and the frontier model race

Anthropic just released Opus 4.6: this puts Anthropic ahead of OpenAI on all the frontier benchmarks, important for them since Opus 4.5 was (slightly) behind GPT5.2 on the leaderboards.

Takeaway? Expect GPT 5.3 to emerge pretty soon.

This is all good for us as end users - the competition pushes similar capability gains whether you use Claude, Gemini, ChatGPT etc. Most businesses will bet on one of the above (for cost, but it also simplifies compliance). The only elephant in the room is Microsoft, which is where most businesses have locked in their AI bet, yet the Microsoft tooling continues to underwhelm.

Yet even if Microsoft is a few months behind the cutting edge, we know that they will get there. So, if I was locked into a Microsoft world, I would be looking at the other tools to see what is coming for me next quarter.