X @Elon Musk

CoolRegi Kusumaatmadja (@abdurrahmanregi):For a trick moment inequality question, it was only GPT 5.2 pro that was able to find the subtlety. If just directly calling the API, gemini 3 pro failed. But now, Grok 4.20 succeeds. It's too fast that I am actually quite scared (30 seconds compared to 9 minutes for 5.2 pro). ...