@future.with.ai98: DeepSeek V4.1 Flash 1017 just dropped 🤯 It's a Flash model, so it's very fast. DeepSeek says it has stronger benchmarks, faster task completion and shorter outputs. It's already on the API, and it's open weights. On DeepSeek's own table, it beats Claude Opus 5 on 4 of 11 benchmarks. All of them are agentic or coding benchmarks. Budget models aren't really supposed to do that. I've already plugged it into my Agent OS. And it built this game while I was talking. If you want to learn AI automation for free, link in the bio #DeepSeek #DeepSeekV41 #OpenWeights #AICoding #AIAgents #OpenSourceAI #AI