A new entry just shook up the Microsoft Foundry AI Model Leaderboard.
Shortly after appearing on the platform, GPT-5.3-Codex moved into the No. 1 spot for Quality, edging out other frontier models by a very slim margin. The leaderboard ranks models across metrics like quality, safety, cost, and throughput, highlighting the trade-offs teams face when choosing a model.
Interestingly, when looking across all categories using a “podium” scoring approach, GPT-5-Nano comes out on top overall thanks to its efficiency and cost performance. Meanwhile, other models lead specific areas like safety and throughput.
When you evaluate models, what matters most in practice: quality, cost, or speed?
Dive into the full article here: Shortly After Debut, GPT-5.3-codex Dominates Microsoft Foundry AI Model Leaderboard - Visual Studio Magazine