lmarena.ai (formerly lmsys.org) on X: “Breaking news from Text-to-Image Arena! 🖼️✨ @GoogleDeepMind’s Imagen 3 debuts at #1, surpassing Recraft-v3 with a remarkable +70-point lead! Congrats to the Google Imagen team for setting a new bar! Try the best text2image at LMArena and cast your vote! More analysis👇 https://t.co/zjD9RvVNUU” / X – https://x.com/lmarena_ai/status/1882164189739073990
“Breaking news from Text-to-Image Arena! 🖼️✨ @GoogleDeepMind’s Imagen 3 debuts at #1, surpassing Recraft-v3 with a remarkable +70-point lead! Congrats to the Google Imagen team for setting a new bar! Try the best text2image at LMArena and cast your vote! More analysis👇
https://x.com/lmarena_ai/status/1882164189739073990
“Inference-time scaling (using more computing power for the AI to “think”) improves the performance of LLMs, as o1 has shown everyone. This paper finds that this same trick also works for diffusion models – commonly used for image generation, among other things.” / X
