RT @LottoLabs: DiffusionGemma 26B-A4B with llama.cpp fork. This is a good example of how diffusion models can process a block of text in parallel as opposed to sequentially
Several AI models have been released or highlighted across various platforms. DiffusionGemma 26B-A4B is noted for its parallel text processing capabilities, while Qwopus 3.6 27b-Coder is now available. Additionally, Hive v0.6 has been released, and there's an opinion that MiniMax, Xiaomi, and DeepSeek models offer a good balance of cost and performance for many use cases. AI
IMPACT Highlights a diverse range of AI model releases and opinions on their value.