DeepSeek v4 Flash experimental with vision running (fast) on an m5 max computer and analyzing an image.
The implementation for Metal / CUDA / ROCm is finished, just doing the last tests today before releasing it.
DeepSeek has released the weights of V4 Flash Vision Exp
Multimodal model for analyzing images, screenshots, interfaces and text. It's closer to Claude Opus 4.8, the company and some testers say
China just dropped a 770 billion parameter AI model.
It’s called Hy4, and Tencent just open-sourced it. But the interesting part isn’t just the number. Hy4 uses a Mixture-of-Experts architecture, meaning it has 770B total parameters but activates only around 49B for each
New episode 🎙️ of #Artificial_Intelligence news
🔹#DeepSeek is close to a $7.4 billion financing round at a valuation of $74 billion 🔹#Anthropic warns of software that stole Claude users’ session cookies and depleted their quota 🔹LEAP 2026 opens in Riyadh with the participation of +200 thousand visitors and announcements from Microsoft and NVIDIA…
You have reached the end of the archive
All of deepseek