Image created with gemini-2.5-flash-image with claude-sonnet-4-5. Image prompt: Photo-realistic Times Square at golden hour completely overtaken by Meta Llama advertising, every billboard showing purple-and-white llama logos, code snippets, and stylized llama mascots, giant illuminated llama sculptures towering over crowds of tech enthusiasts, street vendors selling Llama merchandise, warm amber light reflecting off glass buildings, vibrant open-source energy, shot with cinematic depth of field.
Ollama’s cloud models just got an update! ollama run kimi-k2:1t-cloud Welcome Kimi K2 by @Kimi_Moonshot for the 1 trillion parameter model! ollama run deepseek-v3.1:671b-cloud DeepSeek V3.1 Terminus by @deepseek_ai for the updated updated that is better in language”” / X https://x.com/ollama/status/1971750071483167010
Llama-3-405B seems like an insane model now. Why would you ever create something so unwieldly and not a super sparse MoE?”” / X https://x.com/scaling01/status/1971541644647522595
Results on Test-Time Scaling: Training Llama3.1-8B-Instruct on AIME24, AMC23, MATH500, RESTRAIN delivers +11.0% average gain over TTRL and ETMR. 🧵(4/5) https://x.com/jaseweston/status/1974000970192544248
