English translation is being prepared.
Matched fieldsTitle
←Search results
Search “Video”
English translation is being prepared.
Matched fieldsTitle
English translation is being prepared.
Matched fieldsTitle
English translation is being prepared.
Matched fieldsTitle
English translation is being prepared.
Matched fieldsTitle
English translation is being prepared.
Matched fieldsTitle
Krea launches the all-new Krea Video
On November 6, 2026, Krea released the all-new Krea Video, claiming that cinematic-grade video can be generated directly in the browser. The capabilities listed officially include cinematic camera controls, 100+ popular presets, templates for UGC, as well as video editing, style transfer, and sound generation features, and a trial link has been opened. The product is currently in an open trial phase, with no follow-up development information available yet.
Matched fieldsSummary
NVIDIA adds four open-weight checkpoints to Cosmos 3
Under the Cosmos world model platform, NVIDIA has added four new Cosmos 3 open-weight checkpoints, all post-trained versions of Cosmos 3 Nano, containing 16B trainable parameters and released under the OpenMDW1.1 license, permitting commercial and non-commercial use. The four models respectively target multi-view driving data generation, synchronized streaming simulation of cameras and LiDAR, simulator-controlled generation of photorealistic video, and video prediction from dual-arm robot actions. The original text also presents a four-stage training recipe covering bidirectional adaptation, teacher forcing, consistency distillation, and self-forcing.
Under the Cosmos world model platform, NVIDIA has added four new Cosmos 3 open-weight checkpoints, all post-trained versions of Cosmos 3 Nano, containing 16B trainable parameters and released under the OpenMDW1.1 license, permitting commercial and non-commercial use. The four models respectively target multi-view driving data generation, synchronized streaming simulation of cameras and LiDAR, simulator-controlled generation of photorealistic video, and video prediction from dual-arm robot action
Matched fieldsSummary
TestingCatalog found a Gemini Omni 1.2 Flash reference appearing in GCP
On November 5, 2026, TestingCatalog discovered a reference record for the Gemini Omni 1.2 Flash model on GCP, and stated that it would be a minor upgrade to the then-current Omni 1.1 Flash video generation model. The next day, Google DeepMind launched Gemini Omni 1.2 Flash, which natively supports video generation of up to 15 seconds, with pricing 30% lower than Gemini Omni 1.1, at $0.07 per second for 720p video; the new version enhances reference image adherence and consistency of characters, products, and scenes, supports multi-turn conversational editing, and can replace products or backgrounds without regenerating the entire video, and is fully available in the Gemini API and Gemini for business, with outputs carrying a SynthID watermark. OpenRouter subsequently listed the model, supporting clips of up to 15 seconds, resolutions from 360p to 4K, and native audio, at a per-second price about 30% lower than 1.1.
On November 5, 2026, TestingCatalog discovered a reference record for the Gemini Omni 1.2 Flash model on GCP, and stated that it would be a minor upgrade to the then-current Omni 1.1 Flash video generation model. The next day, Google DeepMind launched Gemini Omni 1.2 Flash, which natively supports video generation of up to 15 seconds, with pricing 30% lower than Gemini Omni 1.1, at $0.07 per second for 720p video; the new version enhances reference image adherence and consistency of characters,
Matched fieldsSummary
Manus Studio supports export to editing tools such as Premiere Pro
On November 5, 2026, Manus announced that Manus Studio projects can now be exported to Premiere Pro, DaVinci Resolve, CapCut, or Final Cut Pro. Users can either create and edit directly with AI in Manus, or export videos to their preferred tools for editing. This news was released by Manus officially, and there have been no further developments so far.
Matched fieldsSummary
Generalist expands access to GEN-1.5 and opens a waitlist
On November 5, 2026, Generalist announced that it was expanding access to GEN-1.5 and opening a waitlist. On the same day, Generalist released a video showing GEN-1.5 completing a thinner O-ring manipulation task, stating that such tasks test the model's dexterity and its ability to perceive millimeter-level details. The post also directed users to read the GEN-1.5 details and join the waitlist. There are currently no further developments.
Matched fieldsSummary
Pangram releases the Image 1.0 and Image Flash 1.0 detection models
On November 4, 2026, Pangram released two computer vision models: Pangram Image 1.0 for detecting AI-generated images, and the smaller and faster Pangram Image Flash 1.0 for frame-by-frame analysis of AI video. According to a report by Unite.AI, the company claims that Image 1.0 can identify 99.2% of AI-generated images with a false positive rate of 0.27%, and achieves 99.96% accuracy on the company's own set of 1,200 synthetic images; Flash 1.0 scores 98.7% on the DF26 benchmark, 29 points higher than the next-best detector. This update also adds heatmap output, native video upload, Chrome extension integration, and image and video API support, with the minimum input reduced to 256×256 pixels; images below 512 pixels come with a low-confidence warning. A report by heise online on the same day noted that the above accuracy figures have not been independently verified.
On November 4, 2026, Pangram released two computer vision models: Pangram Image 1.0 for detecting AI-generated images, and the smaller and faster Pangram Image Flash 1.0 for frame-by-frame analysis of AI video. According to a report by Unite.AI, the company claims that Image 1.0 can identify 99.2% of AI-generated images with a false positive rate of 0.27%, and achieves 99.96% accuracy on the company's own set of 1,200 synthetic images; Flash 1.0 scores 98.7% on the DF26 benchmark, 29 points high
Matched fieldsSummary
PixVerse showcases an action thriller concept film made with its tools
On November 4, 2026, PixVerse released an action thriller concept video, and the post stated that the concept was produced by PixVerse. The video shows a fight breaking out in a luxury hotel; a few minutes later, the female lead appears on the rooftop as a helicopter approaches. Currently there is only this one release information, with no subsequent developments yet.
Matched fieldsSummary
Krea releases Krea 3 technical report
Krea released the Krea 3 technical report on November 4, 2026, providing an in-depth introduction to the data, architecture, and training techniques used by the model. The official Krea AI tweet provides a link to the report, along with a video. There is currently no information on subsequent developments.
Matched fieldsSummary
Hao AI Lab integrates MiniMax H3 Flex into FastVideo
On November 4, 2026, Hao AI Lab announced that it has integrated Pika Labs' open-source MiniMax H3 Flex into FastVideo, claiming inference speeds up to 7 times faster than the official open-source inference code. On a single GB200, generating a 5-second 480p video with audio using the same prompt and random seed took 59.7 seconds with the official code, 26.6 seconds with FastVideo BF16, 16.8 seconds with NVFP4 QAD, and 15.2 seconds with FP4 attention. The official code supports only a single GPU; after FastVideo added multi-GPU support, generating a 15-second video on 4 GPUs was 7 times faster; generating a 5-second 480p first-and-last-frame video on a single DGX Spark took 59.5 seconds. The project also releases the weights, inference code, and complete training recipe, and supports the 2K model of Open-H3-Regenerate-2K.
On November 4, 2026, Hao AI Lab announced that it has integrated Pika Labs' open-source MiniMax H3 Flex into FastVideo, claiming inference speeds up to 7 times faster than the official open-source inference code. On a single GB200, generating a 5-second 480p video with audio using the same prompt and random seed took 59.7 seconds with the official code, 26.6 seconds with FastVideo BF16, 16.8 seconds with NVFP4 QAD, and 15.2 seconds with FP4 attention. The official code supports only a single GPU
Matched fieldsSummary