Google’s Gemini 4 Argon has only been available to developers for a day, yet people are already using it to create games, build 3D environments, and even teach a real robot new skills. 🤯 Argon is the first model in the Gemini 4 lineup, designed to handle complex, multi-step tasks involving coding, research, and financial analysis. It can also generate up to 1 million tokens in a single output. Developers gained access on October 13 through the paid Gemini API in Google AI Studio and Android Studio. Google AI Ultra subscribers can also experiment with it through AI Studio and Google Antigravity. The model has also claimed the top position among 48 models on the Vals Index, an independent benchmark. Its first 24 hours have already produced some impressive experiments. Developers transformed photos of desks into racing games, reportedly built a Minecraft-style clone in 19 minutes for around $3, created an elaborate temple environment in Blender, and trained a humanoid robot to play c...