1Qwen 3.5: small models with impressive performance (opens in new tab)(twitter.com)6moondistance23d ago0
2Language Models Are Injective and Hence Invertible (opens in new tab)(arxiv.org)1moondistance4mo ago2
7DeepSeek-R1 at 3,872 tokens / second on a single Nvidia HGX H200 (opens in new tab)(blogs.nvidia.com)13moondistance1y ago1
8ByteDance Doubao-1.5-pro matches GPT 4o benchmarks at 50x cheaper (opens in new tab)(twitter.com)4moondistance1y ago0
9US-China Commission top recommendation: Manhattan project for race to AGI [pdf] (opens in new tab)(uscc.gov)4moondistance1y ago0
10AI scans RNA 'dark matter' and uncovers 70k new viruses (opens in new tab)(nature.com)1moondistance1y ago0
11Llama 405B 506 tokens/second on an H200 (opens in new tab)(developer.nvidia.com)21moondistance1y ago5
12SenseNova 5.5 claims SOTA LLM benchmark results (opens in new tab)(twitter.com)2moondistance1y ago0
13New AI Training Technique Is Drastically Faster, Says Google (opens in new tab)(decrypt.co)84moondistance1y ago38
14Nvidia open source LLM Nemotron 4 340B at top of the charts [pdf] (opens in new tab)(d1qx31qr3h6wln.cloudfront.net)17moondistance1y ago1
