ChengsenWang/ChatTime-1-7B-Base · Hugging Face
https://huggingface.co/ChengsenWang/ChatTime-1-7B-Base
“Great Open-source model and paper from ByteDance Face dubbing made simple – from sound to lips in one smooth move. Skip the motion middleman, direct audio-to-lip sync in latent space eliminates the middleman for better expressions. LatentSync introduces end-to-end lip sync
https://x.com/rohanpaul_ai/status/1876213656645763098
“🚀 Exciting News! We’re thrilled to announce the launch of Qwen Chat (
https://x.com/Alibaba_Qwen/status/1877426465349972113
“Today, we’re launching early access for North! Our all-in-one secure AI workspace platform combines LLMs, search, and agents into an intuitive interface that effortlessly integrates AI into your daily work to achieve peak productivity.
https://x.com/cohere/status/1877335657908949189
“NEW: DeepSeek-V3 is now available on Together AI 🚀 – Available on Together AI with the full 131K context & opt-out privacy controls – Ranked #7 in Chatbot Arena – the only open model in the top 10! – An efficient giant: 671B MoE params but only 37B activated Props to
https://x.com/togethercompute/status/1876307295405248531
“Process Reinforcement through Implicit Rewards We present PRIME (Process Reinforcement through IMplicit REwards), an open-source solution for online RL with process rewards, to advance reasoning abilities of language models beyond imitation or distillation. With PRIME, starting
https://x.com/_akhaliq/status/1875039314771660989
North: The AI Platform Where Work Gets Done | Cohere
https://cohere.com/north
EXAONE-3.5 – a LGAI-EXAONE Collection
https://huggingface.co/collections/LGAI-EXAONE/exaone-35-674d0e1bb3dcd2ab6f39dbb4
“A simple explanation of how @ai16zdao’s eliza framework works 🧵:”
https://x.com/0xCygaar/status/1874575841763770492
“@lalopenguin ollama run phi4”
https://x.com/ollama/status/1877058716069147092
Introducing Sonus-1: A New Era of LLMs – Goran Babarogic Product Designer
https://sonus.ai/blog/sonus-1
Models Table – Dr Alan D. Thompson – LifeArchitect.ai
Nous DisTrO
https://distro.nousresearch.com/
PleIAs/Pleias-3b-Preview · Hugging Face
https://huggingface.co/PleIAs/Pleias-3b-Preview
They Said It Couldn’t Be Done
https://huggingface.co/blog/Pclanglais/common-models
“Updated the list with more models – @NousResearch Hermes, IBM Granite, etc 🔥”
https://x.com/reach_vb/status/1874829731117572150
“evals look very very interesting”
https://x.com/reach_vb/status/1874869340111335783
DeepSeek
“We’ve added DeepSeek-V3 to the @aisdk playground. It pretty consistently identifies as ChatGPT & OpenAI 🤨
https://x.com/rauchg/status/1875627666113740892
Deepseek: The Quiet Giant Leading China’s AI Race
https://www.chinatalk.media/p/deepseek-ceo-interview-with-chinas
“One of the implications of China’s DeepSeek 3 being competitive with frontier models and open (and coming out so quickly after frontier US models) is that it seems that US (let alone EU) regulation is increasingly unlikely to stop continued AI model development.” / X
https://x.com/emollick/status/1875279332048039970
Grok
xAI’s next-gen AI model didn’t arrive on time, adding to a trend | TechCrunch
xAI’s next-gen AI model didn’t arrive on time, adding to a trend
Grok may soon get an ‘Unhinged Mode’ | TechCrunch
“@Scobleizer Yeah, Grok will be better at understanding complex group speech in a noisy environment than people are” / X
https://x.com/elonmusk/status/1876730201264062480
“BREAKING: Elon Musk has just confirmed that Grok 3 will be released in 3-4 weeks.
https://x.com/cb_doge/status/1876703333156614232
“did i not tell you grok 3 is coming. by far the best model i’ve ever used and its not even close.” / X
https://x.com/iruletheworldmo/status/1876728550864077045
NVIDIA
“Introducing NVIDIA Cosmos, an open-source, open-weight Video World Model. It’s trained on 20M hours of videos and weighs from 4B to 14B. Cosmos offers two flavors: diffusion (continuous tokens) and autoregressive (discrete tokens); and two generation modes: text->video and
https://x.com/DrJimFan/status/1876516972512559170
NVIDIA Cosmos video dataset – Dr Alan D. Thompson – LifeArchitect.ai
Cosmos World Foundation Model Platform for Physical AI | Research
https://research.nvidia.com/publication/2025-01_cosmos-world-foundation-model-platform-physical-ai
Cosmos – a nvidia Collection
https://huggingface.co/collections/nvidia/cosmos-6751e884dc10e013a0a0d8e6
Meta/Llama
ruliad/deepthought-8b-llama-v0.01-alpha · Hugging Face
https://huggingface.co/ruliad/deepthought-8b-llama-v0.01-alpha
llama-models/models/llama3_3/MODEL_CARD.md at main · meta-llama/llama-models
https://github.com/meta-llama/llama-models/blob/main/models/llama3_3/MODEL_CARD.md
meta-llama/Llama-3.3-70B-Instruct · Hugging Face
https://huggingface.co/meta-llama/Llama-3.3-70B-Instruct
“It’s quite a bit crazy that one fine day @ggerganov released llama.cpp and it set an entire new path for bringing intelligence closer to the edge! 🤯 2025 is all about bringing the frontier closer to everyone and beyond
https://x.com/reach_vb/status/1875595202683134009
“Introducing Trend Finder 2.0 🔦 Have AI monitor trending topics on social media & websites then instantly send you slack alerts. Now featuring: – Llama 3.1 for better speed and cost via @togethercompute – Web page support via @firecrawl_dev – Streamlined setup Check it out:
https://x.com/ericciarla/status/1874860989008785542
Mark Zuckerberg gave Meta’s Llama team the OK to train on copyrighted works, filing claims | TechCrunch
Mark Zuckerberg gave Meta’s Llama team the OK to train on copyrighted works, filing claims
Meta Secretly Trained Its AI on a Notorious Piracy Database, Newly Unredacted Court Docs Reveal | WIRED
https://www.wired.com/story/new-documents-unredacted-meta-copyright-ai-lawsuit/
Mistral
“1-year at @MistralAI. Full of gratitude. A year ago, I took a leap of faith into the unknown, excited and slightly nervous, moving to Paris and beginning my journey at Mistral. I remember arriving with a group of others and being warmly greeted by one of the kindest coworkers.
https://x.com/sophiamyang/status/1875219788407980237
Phi
“The model weights of Microsoft’s phi-4 14B LLM are now finally publicly available. Thanks to a kind community contribution, it’s now also available in LitGPT already (litgpt finetune microsoft/phi-4):
https://x.com/rasbt/status/1877380409387802946
Microsoft makes Phi-4 model fully open source on Hugging Face | VentureBeat
Microsoft makes powerful Phi-4 model fully open-source on Hugging Face
microsoft/phi-4 · Hugging Face
https://huggingface.co/microsoft/phi-4
Qwen
“Qwen-Agent Agent framework and applications built upon Qwen>=2.0, featuring Function Calling, Code Interpreter, RAG, and Chrome extension. pip install -U “qwen-agent[gui,rag,code_interpreter,python_executor]” # [gui] for Gradio-based GUI support; # [rag] for RAG support; #
https://x.com/_akhaliq/status/1874574960179704246
“Qwen 32B (4-bit) generates at >40 toks/sec on an M4 Max with assisted decoding and Qwen 0.5B as the draft model. Coming soon to mlx-lm. Compare regular decoding (left) to assisted decoding (right):
