ai-models· 4 min read · Jun 17 Gemma 4 12B: Google's Encoder-Free Multimodal Model Fits in 16GB of RAM
ai-models· 3 min read · May 22 DeepSeek Raises $10.3B and Cuts Prices 75%: The Math Is Starting to Scare Everyone
ai-models· 4 min read · May 17 MTP Support Lands in llama.cpp Main: Up to 2.4x Faster Token Generation