If the initial benchmarks scores (and graphs used for PR) showcased a 3x reduction in size for the same performance, I think the broader public reception would have been less tepid. Just looking at this, it just seems 7% behind other existing models in the 4.x series…
INNOVATION
-

Hugging Face Releases Transformers Version 5 RC
By
–
Today we release the transformers version 5 RC! With this, we enable e2e interoperability with our friends in ecosystem, ease up adding new models and simplify the library Read our blog to learn more: https://
huggingface.co/blog/transform
ers-v5
… -
Kling AI Unveils Multimodal Kling O1 Model
By
–
Kling AI released the Kling O1 model with multimodal understanding that can unify input across texts, images, and videos!
— 🚨 AI News | TestingCatalog (@testingcatalog) 1 décembre 2025
AI video week 🔥 https://t.co/r7WXgVtLRf pic.twitter.com/HtXMzX5ZuhKling AI released the Kling O1 model with multimodal understanding that can unify input across texts, images, and videos! AI video week
-

DeepSeek-V3.2 outperforms GPT-5 on benchmarks after release
By
–
DeepSeek-V3.2 dropped last night and absolutely crushes GPT-5 on benchmarks
-

Whisper Thunder Emerges as Google Competitor in AI Market
By
–
whisper thunder reveals itself. looks pretty good. i'm glad google has competition here, was worried they are running away too fast!!! https://t.co/QkQ0vFn3lg
— Soumith Chintala (@soumithchintala) 1 décembre 2025whisper thunder reveals itself. looks pretty good. i'm glad google has competition here, was worried they are running away too fast!!!
-
Post-training bottlenecks solved through refined methods and data
By
–
"The lesson is post-training bottlenecks are solved by refining methods and data" Zhibin Gou (@zebgou) If Gemini-3 proved continual scaling pretraining, DeepSeek-V3.2-Speciale proves scaling RL with large context. We spent a year pushing DeepSeek-V3 to its limits. The lesson is post-training bottlenecks are solved by refining methods and data, not just waiting for a better base. — https://nitter.net/zebgou/status/1995462720078934213#m
-

UMA Robotics Startup Launched by Google and HuggingFace Veterans
By
–
Congrats to @psermanet
, @RemiCadene and team on starting UMA – https://
uma.bot . Considering Pierre's track record at Google Robotics, and Remi's phenomenal work on LeRobot at HuggingFace, this is going to be great. I'm so glad to be an angel investor. -

Meta at NeurIPS 2025 San Diego: Latest AI Research Demos
By
–
We’re in San Diego this week for #NeurIPS2025! Stop by the Meta booth (#1223) to meet our team and check out: Demos of our latest research including DINOv3 and UMA Lightning talks from researchers behind SAM 3, Omnilingual ASR and more (see schedule below) Hands-on
-

Ultimate CES 2026 Guide: AI Panels, Parties and Robots
By
–
💥The Ultimate CES 2026 Guide: Hottest AI Panels, Parties And Robots By @Forbes v/ @chidambara09 #CES2026 #Innovation #AI #WearableTech #Robotics #Engineering #HealthTech forbes.com/sites/martinepari… Chidambara .ML. (@chidambara09) forbes.com/sites/martinepari… my @enilev🪔mom Mr @BetaMoroney sir ThX🌐#BOSCh #ecb #Pfizer #Chevron #CES2026 @CES #ces #usa #uk #europe #frenchtech #Technology #DiGitaLhealth #ai Chidu09.BOSCh.com 🍁 — https://nitter.net/chidambara09/status/1995158047237546319#m
→ View original post on X — @marcusborba, 2025-12-01 16:32 UTC
-

Axelera AI Showcases Edge AI Performance at CES 2026
By
–
CES 2026 is just around the corner and Axelera AI is ready to show you what true edge AI performance looks like. Join us at the Venetian Tower (Suite 29-111) to experience: Real-time 8K video inference AI motion tracking for your golf swing Axelera AI Co-Pilot:
