传 Gemini 4 Flash 在 LMArena 达约 300 tokens/秒

Rajath Gowda @buildwithrajath原推发布于 2026.10.28

在 X 上打开原推

🚨 Gemini 4 Flash might be hiding one of Google's biggest competitive advantages yet. It's reportedly hitting ~300 tokens/sec on LMArena. Now imagine getting Gemini 4 Argon-level intelligence at that speed, for the price of a Flash model. It would be fucking INSANE. If Google can pull that off, it won't just have a faster model. It'll have a model that's smarter, faster, AND cheaper than what competitors can offer. OpenAI and Anthropic, what the fuck is the counter to that?