The Decoder6565Runway has released the open-source robot control model Praxis-1, which learns to control robots through video and has shown excellent performance in its own tests, scoring 46.2% on the Nvidia RoboLab-120 evaluation, leading similar models. The model requires additional training to adapt to new robot systems, the standard version requires an 80GB memory GPU, and it has been made available on Hugging Face.
MarkTechPostThe Elo ranking shows Alibaba Wan 3.0 (1175 T2V Elo) at the top with a price of $0.20 per second; Utopai X (1170 T2V Elo) is second but has no API; MiniMax H3 (1161 T2V Elo) is the first open-source video model with a price of $0.08 per second, but it is regionally restricted. The ranking is based on blind human preference tests, combined with the price per second of the API and the status of weighted licenses.
Ant Group inclusionAI · Hugging Face new modelsinclusionAI has released the Duo 1.2 causal-base model, a full-duplex multimodal world model for real-time audio-video interaction, featuring improved generation quality, and model weights can be downloaded from Hugging Face.
Ant Group inclusionAI · Hugging Face new modelsinclusionAI releases Duo-1.0-distill, a full-duplex multimodal world model for real-time audio-video interaction, including Realization weights with four-step distillation, which needs to be used in conjunction with the causal-base suite.
Hugging Face Daily Papers6565PhysLoop utilizes large language models to guide video generation, selection, and correction, achieving physically consistent self-improvement and controllable data construction. The framework drives recursive self-improvement (RSI) of video generation models with limited human intervention by identifying and correcting physical errors in generated videos, and releases a diverse dataset containing 100K videos along with generation tools.
Hugging Face Daily PapersBirdDreams jointly generates bird's-eye view (BEV) videos and bird's-eye view videos, making scene layout a part of predicting the future. By providing an overhead view of vehicle position and movement through BEV, it reduces occlusion, while the perspective video retains scene details. On PhysicalAI-AV and nuScenes, BirdDreams more faithfully follows the input trajectory of its own vehicle and generates more reasonable self-motion without HD map or vehicle state input.
The DecoderAnthropic has released the Claude Meetings feature, allowing Claude to join video conferences as a visible participant, answer questions, generate documents and whiteboards, take notes and action items, and respond in text format by default. With authorization, it can access connected tools. After the meeting, all notes, text records, and documents are uploaded to the dedicated claude.ai page. Privacy controls default to preventing Claude from joining meetings with external participants.
Adobe announced four AI experimental projects: Project Screen Test designs AI characters through dialogue; Project Sight Span converts photos into explorable 3D scenes; Project Sound Bender allows free adjustment of video sound effects; Project Rinse Repeat automatically replicates editing operations.
The Verge · AISame story · 2 outletsAdobe has launched experimental AI tools Project Screen Test to design AI characters through dialogue and generate audition animations; Project Sight Span converts photos into explorable 3D environments; Project Sound Bender allows users to customize sound effects like a color wheel; Project Rinse Repeat can replicate edits and apply them automatically to other objects in the same image
Adobe has introduced the 'Re-light' tool in Photoshop, which allows adding and adjusting the position, color, and brightness of light sources, and modifications do not damage the original image; the upgraded 'Remove' tool can eliminate shadows, reflections, footprints, and ripples left by objects and reconstruct the background.
Ars Technica · AI6565Amazon introduces Ring Flight drone and Ring Rover Cam robot, offering 2K video features with 148-degree field of view and 360-degree rotation, respectively. The drone has a flight time of 8 minutes, and the robot is priced at $1000, supporting 4K recording and cloud storage for 180 days. Both are controlled via Alexa Home Map, but the drone's performance decreases on reflective surfaces, and the robot is limited to single-story indoor use.
The Verge · AIAdobe announced new features for Premiere, Photoshop, and Lightroom at its Max event, including automatic object removal, flaw removal, and elimination of other unwanted distractions, while also launching new generative AI tools and cross-Adobe application collaboration experiences.
AWS Machine Learning Blog5858NVIDIA Nemotron 3.5 Super VL is now available on Amazon SageMaker JumpStart, supporting text generation from image, video, and text inputs.
Microsoft filed a patent application, using AI to almost instantly modify live broadcast content and synchronize mouth movements, capable of replacing verbal fillers, stuttering, offensive words, etc., supporting real-time translation, with a delay of 1 to 10 seconds, potential application scenarios include lectures, sports event broadcasts, etc., currently only a patent application, the technology has not yet been authorized.
Google has launched the AI video generation model Gemini Omni 1.2 Flash, natively supporting the longest 15-second video, with a duration limit increased by 50% compared to the previous generation's 10 seconds. Users can set the generation duration from 3 seconds to 15 seconds in 1-second increments.
Dream Short Drama Studio provides a complete creative process from script to storyboard, supporting one-sentence inspiration or existing script creation, generating a 10-episode outline and the first three episodes of the script with clear annotations of character actions and expressions. Seedance 2.5 offers low costs as low as 0.15 yuan per second, supports smart preview and 3D scene verification, and the institutional partner program provides IP connection and publicity support, with annual packages for individuals as low as a 4.5 discount.
Alibaba's new video generation model Wan3.0, through consistency, realistic texture, style mastery, and video editing capabilities, achieved stable reproduction of character, product, and spatial relationships in Taobao short dramas 'Nine-Tailed God's Arrival,' 'Clumsy Hands Clumsy Feet Restaurant,' and 'Beyond the Final Chapter,' assisting in brand integration and promoting transaction conversion, demonstrating the commercial value of AI video models in the e-commerce field.
MarkTechPostFeaturedSame story · 6 outlets7272Google releases video generation and editing model Gemini Omni 1.2 Flash, natively supporting up to 15-second clips, with 720p video pricing at $0.07 per second, a reduction of about 30% compared to Omni 1.1 Flash
AMD Next-Gen Mobile APU Medusa Point Scores 1560 Single-Core, 11575 Multi-Core in Lenovo Test System with Geekbench 7, Equipped with 10 Cores 20 Threads (4 Zen 6 Performance Cores + 6 Zen 6c Efficiency Cores), 16MB L3 Cache, Supports Instruction Sets like AVX-512
Google launched the video generation model Gemini Omni 1.2 Flash on November 5 local time, and fully opened it in Gemini API and Gemini for Business. The model introduces native 15-second video creation, multi-round conversational editing, and enhanced reference following capabilities, with a 30% price reduction compared to Gemini Omni 1.1, priced at $0.07 per second in 720p resolution.
Google has launched the Gemini Omni 1.2 Flash video generation model and made it available on the Gemini API and the Gemini for business platform. The new native version supports videos up to 15 seconds long, with customizable durations from 3 to 15 seconds in 1-second intervals. The 720p generation price is $0.07 per second, a 30% reduction compared to the 1.1 version.
TechCrunch · AI6565Meta announces that its consumer AI agent Muse's voice and video chat features Muse Realtime Avatar and Muse Realtime Voice are now available to all users in the United States. Users can engage in real-time conversations by interacting with the cute animated avatar Jolly, which can be customized according to user preferences and uses subtle expressions and gestures that match the mood of the conversation.
Hugging Face BlogFeatured7171NVIDIA has added four new open-source weight checkpoints for Cosmos 3 under the Cosmos World Model platform, all of which are post-trained versions of Cosmos 3 Nano, containing 16B trainable parameters, using the OpenMDW1.1 license, allowing commercial and non-commercial use.
Why it matters: The original text provides four open-source weight checkpoints and a four-stage training recipe, allowing readers to judge how to adapt the world model to their own sensors and robot bodies.
The DecoderFeaturedSame story · 6 outlets9292Google DeepMind has made the Gemini Omni 1.2 Flash video generation model generally available, extending clip duration from a maximum of 10 seconds to 3 to 15 seconds, and reducing the price of 720p video to 0.07 USD per second, a 30% decrease from the previous generation.
Google AI BlogFeaturedSame story · 6 outlets7878Google has released Gemini Omni 1.2 Flash and has officially opened it through the Gemini API and Gemini for business. The duration of native video generation has increased from 10 seconds in Omni 1.1 to 15 seconds, with a step length of 1 second specified from 3 to 15 seconds. The price of 720p has been reduced to 0.07 US dollars per second, a decrease of 30% from 1.1.
Google DeepMindFeaturedSame story · 6 outlets7878Google DeepMind has released Gemini Omni 1.2 Flash, increasing the native video generation duration from 10 seconds in Omni 1.1 to 15 seconds, and it has been fully opened in the Gemini API and Gemini for business.
Why it matters: The original text lists specific changes in duration, price, and reference consistency, allowing readers to judge how the cost and workflow of video generation will change.
Hugging Face Daily Papers6565EmoReward adapts multi-reward weight allocation by extracting evidence reliability from reward trajectories collected during actual training runs to calibrate reward weights for emotional video generation. This mechanism first evaluates specific grounding cues based on emotional prompts, then scores videos during the generation phase using heterogeneous signal primitives, and calibrates their contributions through reflective reward gating.
Hugging Face Daily Papers6565The TurnWise framework improves streaming video understanding by round-level credit assignment and decision weight balancing. Its round-level credit assignment constructs group relative advantages using prefix scores from silent completions, and round-mean policy optimizes the token loss per round. Experiments show this method outperforms existing streaming video understanding RL methods.
Xiaomi REDMI K100 Phone Adds 3D Camera Movement AI Feature, Select Static Photos from Album to Generate Wide-Angle 3D Camera Movement Videos, Lens Can Orbit, Pull, Tilt, and Keep Character Appearance Consistent. This Feature is Based on Xiaomi's Visual Generation Framework WorldCast, Interpreting Depth Structure of the Scene to Restore Flat Photos into 3D Scenes and Virtual Camera Path. Xiaomi Claims WorldCast is a Breakthrough in Controllable Visual Generation Direction, Will Explore More Complex Camera Movement Forms and 4D Video Generation Technology.
Xiaomi Smart Camera 4 Pro's launch price was 329 yuan, reduced to 252 yuan today after JD.com's government subsidy. It now supports physical call buttons for one-click WeChat calls, enabling two-way calls between the camera and WeChat on smartphones or the Mi Home app.
Orca Culture Group's 'Jingrui AI', the industry's first AI film production and management platform, officially launched, integrating over 50 mainstream AI video generation models and over 10 self-developed models, saving more than 30% in AI video costs per minute compared to traditional methods through precise control and process optimization. The platform's official website (www.jingruiai.cn) opens for full registration, team seats unlimited in number, no membership subscription required.
Hugging Face Daily Papers3838DreamX uses pre-trained video diffusion models to learn compact, appearance-invariant latent actions, achieving cross-subject transfer and efficient policy learning. DreamX extracts latent actions from the rich spatiotemporal features of the video model, compressing each frame 300 times (1200→4 tokens).
Hugging Face Daily Papers6565MoviEdit framework, through cascaded bidirectional autoregressive modeling and new benchmark tests, maintains cross-scene coherence while reducing manual data annotation, significantly outperforming existing open-source models with performance comparable to industry leading models. The framework automatically generates 87K pairs of paired editing videos for cross-shot supervision, establishes a global editing draft using a lightweight bidirectional stage, and then performs autoregressive propagation and refinement.
