Today's experiment DeepSeek-V4.1-Flash 347 GB, running Q3_K_M with 72gb vram
It's slow at around 18 tps, but it's a bit smart. Tomorrow, I'll do some more tuning and play around with making it into the mid 20s.
Giving away GPT-6 Astra, Claude Fable 5.1 and 20+ frontier AI models - just by playing a game
Rabbit Hop just went live on freeLLM Top players win real API keys every day: - GPT-6 Astra - Claude Fable 5.1 - DeepSeek V4 Pro - Qwen 3.8 Max - Kimi K3 - and other frontier models
AI agents might have just become a lot cheaper.
DeepSeek V4.1 Flash is available through Token Harbor with a free API you can plug into your Agentic OS. We’re using it across AI workflows to build and automate for free. Comment “Agent OS” for the setup.
ChatGPT me DeepSeek me?
Which is the real winner! I explained it in the short video — its application is simpler than it seems. Where are you stuck on this issue?
You have reached the end of the archive
All of deepseek