Radar

DeepSeek AI news

New AI models and products, prices and deals from DeepSeek. Newest first.

Importance 2 of 5r/LocalLLaMA

Update: tandem.cpp — DeepSeek-V4.1 on a 64 GB PC with a 16 GB card, and AMD (ROCm) support

Quick update on tandem.cpp, my patch set for llama.cpp that runs MoE models too big for your VRAM faster. It keeps the hot experts on the GPU and lets the CPU help with the rest. Repo: https://github.com/ewhacc/tandem.cpp DeepSeek-V4.1-Flash IQ2_XS (ISTA GSQ-RCO, 139 GiB of experts) on normal PCs Decode tok/s, one…

Importance 3 of 5r/LocalLLaMA

How long are we gonna depend bunch of model creators (like Qwen, Gemma, NVIDIA, etc.,) for ~100B models?

I think we're fine till 2027 with models(~100B .... Qwen 4.X & 5, Gemma-5, NVIDIA would release some Medium/Big size 25-200B models). After that it's gonna be critical, as many model creators are increased/increasing model size to 300B+. Future trend probably gonna be around 500B. That size range is not suitable for…