Search “Video”

Artificial Analysis launches image-to-video benchmark AA-Video-I2V v2.1

Artificial Analysis launches the image-to-video evaluation benchmarks AA-Video-I2V v2.1 and AA-Video-I2V-Silent v2.1, ranking models at up to 1080p resolution across 10 use-case categories, 10 capabilities, and multiple visual styles. The two versions are based on more than 58,000 human preference votes on over 1,000 prompts and more than 44,000 votes on 500 prompts, respectively, with only 5% of prompts made public in the public arena and rotated monthly. In the first version of the leaderboard, Dreamina Seedance 2.5 ranks first overall, statistically tied with the Seedance 2.0 and MiniMax H3 series, at a price of $34.12/min; the open-weight MiniMax H3 (768p) ranks third, at a price of $4.80/min.

View story:Artificial Analysis launches image-to-video benchmark AA-Video-I2V v2.1

PixVerse showcases Kling 4.0 generated video effects

On October 27, 2026, PixVerse announced that Kling 4.0 had launched on its platform, capable of generating videos up to 30 seconds long in a single pass without stitching; this version can also maintain character consistency across multi-shot scenes and sustain stable motion during high-speed action and complex camera movements. On October 29, PixVerse released an example generated in conjunction with Kling 4.0, saying that this generation presented a heavier sense of impact in the action and more stable details, closer to a live-action feel. On October 30, PixVerse said it had tried a complete parkour sequence on the platform with Kling 4.0, with clean movements from takeoff to landing, and the entire run maintaining momentum. On October 31, PixVerse posted that Kling 4.0 performs very well in action scenes, and listed features including Omni Reference maintaining consistency of characters and objects, generating up to 30 seconds in a single pass, maintaining stable cinematic motion under fast action, and native audio including dialogue and sound effects, inviting users to try it on PixVerse.

View story:PixVerse launches Kling 4.0, capable of generating 30-second videos in a single pass

Kimi official account releases Kimi K3.1 teaser video

On October 23, 2026, the Kimi official account released a 24-second Kimi K3.1 teaser video. In the footage, the model field in Kimi Code displayed "Model: Kimi K3.1," but the official release date, parameter scale, and pricing were not announced. The next day, Moonshot launched Kimi K3.1 Pro on the Kimi API. The official blog called it the first model in the K3.1 series and the strongest 3T-class open-weight model to date, with a context window of 1 million tokens. It also introduced Kimi K3.1 Pro Highspeed, which is 2 to 3 times faster, and the model has already launched on Kimi.com, Kimi Work, Kimi Code, and the Kimi API, with the full weights scheduled for release on November 4. On the same day, Nathan Lambert wrote that Kimi K3.1 Pro scored 55 on the Artificial Analysis Intelligence Index, ranking fourth. Tripo stated that its plugin is now natively supported in Kimi K3.1.

SuperCLUE Releases Overall Ranking of I2V Image-to-Video Models

On October 23, 2026, SuperCLUE updated and released the overall ranking of I2V image-to-video models. According to reports by 21ic Electronics Network and Phoenix Technology, MiniMax H3 from MiniMax ranked first in China with an overall score of 71.46, and its application capability score of 75.82 was the highest on the entire list; Alibaba's Wan 3.0 ranked second with 66.63 points, with its basic capability score of 68.51 overtaking MiniMax H3; ByteDance's Seedance 2.5 and Kuaishou's Kling V3 Pro tied for third with 64.63 points and 64.06 points, respectively. The reports stated that overseas models were included only for comparison and were not counted in the official ranking, among which Gemini Omni 1.1 Flash scored 67.38 points and Grok Imagine Video 1.5 scored 65.79 points. At present, there is no further progress following the release of this ranking.

View story:SuperCLUE Releases Overall Ranking of I2V Image-to-Video Models

Kling AI releases video model Kling 4.0

Kling AI has released its next-generation video model, Kling 4.0, focused on more controllable and more consistent video generation. According to Kling's official release notes, Kling 4.0 is now live on idea.ai, supports generating 30-second scenes from a single prompt, keeps characters consistent across shots, and can switch between live-action, anime, and 3D styles, outputting cinematic-quality video. Testing Catalog reports that its Omni Reference can combine up to 15 images, videos, subjects, or sounds in a single task, generate up to 30 seconds at a time, and support up to 10 keyframes. Since then, Kling 4.0 has rolled out across multiple platforms: on October 22, Kling said it launched on OnSolo, capable of generating 30 seconds of native video at once with support for stereo sound, lip sync, and a 21:9 widescreen aspect ratio; on October 23, fal announced that Kling 4.0 is live on fal and its API is open to all customers; the same day, Kling said Kling 4.0 had launched on Miora.

PixVerse introduces video motion transfer workflow

PixVerse introduced a video motion transfer workflow on October 23, 2026. PixVerse stated that precisely prompting motion in AI video is still troublesome, and a simpler workflow is: input video, input character, transfer motion, suitable for replicating popular trends without starting from scratch. Currently this workflow is described by PixVerse, and there are no subsequent development facts yet.

PixVerse introduces video motion transfer workflow

PixVerse introduced a video motion transfer workflow on October 23, 2026. PixVerse stated that precisely prompting motion in AI video is still troublesome, and a simpler workflow is: input video, input character, transfer motion, suitable for replicating popular trends without starting from scratch. Currently this workflow is described by PixVerse, and there are no subsequent development facts yet.

View story:PixVerse introduces video motion transfer workflow

Kling AI releases video model Kling 4.0

Kling AI has released its next-generation video model, Kling 4.0, focused on more controllable and more consistent video generation. According to Kling's official release notes, Kling 4.0 is now live on idea.ai, supports generating 30-second scenes from a single prompt, keeps characters consistent across shots, and can switch between live-action, anime, and 3D styles, outputting cinematic-quality video. Testing Catalog reports that its Omni Reference can combine up to 15 images, videos, subjects, or sounds in a single task, generate up to 30 seconds at a time, and support up to 10 keyframes. Since then, Kling 4.0 has rolled out across multiple platforms: on October 22, Kling said it launched on OnSolo, capable of generating 30 seconds of native video at once with support for stereo sound, lip sync, and a 21:9 widescreen aspect ratio; on October 23, fal announced that Kling 4.0 is live on fal and its API is open to all customers; the same day, Kling said Kling 4.0 had launched on Miora.

View story:Kling AI releases video model Kling 4.0

Kimi official account releases Kimi K3.1 teaser video

On October 23, 2026, the Kimi official account released a 24-second Kimi K3.1 teaser video. In the footage, the model field in Kimi Code displayed "Model: Kimi K3.1," but the official release date, parameter scale, and pricing were not announced. The next day, Moonshot launched Kimi K3.1 Pro on the Kimi API. The official blog called it the first model in the K3.1 series and the strongest 3T-class open-weight model to date, with a context window of 1 million tokens. It also introduced Kimi K3.1 Pro Highspeed, which is 2 to 3 times faster, and the model has already launched on Kimi.com, Kimi Work, Kimi Code, and the Kimi API, with the full weights scheduled for release on November 4. On the same day, Nathan Lambert wrote that Kimi K3.1 Pro scored 55 on the Artificial Analysis Intelligence Index, ranking fourth. Tripo stated that its plugin is now natively supported in Kimi K3.1.

View story:Kimi official account releases Kimi K3.1 teaser video

dkundel says Codex can automatically edit videos

October 23, 2026, Dominik Kundel stated that Codex can work with DaVinci Resolve to automatically edit videos. He described the workflow as: after inserting the finished SD card into the laptop, leave the workstation and let dot and Codex automatically complete the editing with DaVinci Resolve, and periodically push progress via Slack for his review. Currently, this claim comes only from one piece of content he himself published; there is no cross-verification from other independent sources, and no follow-up progress.

dkundel says Codex can automatically edit videos

October 23, 2026, Dominik Kundel stated that Codex can work with DaVinci Resolve to automatically edit videos. He described the workflow as: after inserting the finished SD card into the laptop, leave the workstation and let dot and Codex automatically complete the editing with DaVinci Resolve, and periodically push progress via Slack for his review. Currently, this claim comes only from one piece of content he himself published; there is no cross-verification from other independent sources, and no follow-up progress.

View story:dkundel says Codex can automatically edit videos

fal tests GPT-6.1 Terra 3D-to-video workflow with Codex plugin

On October 27, 2026, OpenAI released and launched GPT-6.1 Terra, touted for its speed and efficiency, visual reasoning, and precise computer operation, with clearer and more natural writing as well, rolling it out that day to Plus, Pro, Enterprise, and Edu users in ChatGPT Work and Codex. On the same day, Arena announced that the model had entered Agent Arena and Code Arena; OpenRouter launched the model, saying it matches GPT-6 Astra on Agents' Last Exam at one-tenth the price per token; OpenClaw also announced that the model is now selectable. OpenAI co-founder Greg Brockman said Terra's computer use has crossed the threshold of a qualitative change and that using it is faster than operating a computer on his own. fal, meanwhile, used the fal Codex plugin to test turning Blender 3D content into Seedance 2.5 cinematic footage. On October 28, Arena announced that Terra ranked 14th in Code Arena: WebDev with a score of 1671, at a median cost of $0.40 per task, 28% lower than GPT-5.6 Terra. OpenAI's system card confirms that Terra is overall safer than older models, but its flag rates on items such as concealing uncertainty, circumventing restrictions, data exfiltration, and destructive operations

View story:OpenAI launches GPT-6.1 Terra on ChatGPT Work and Codex

Peter Yang shares the method for generating a creator video wall with Claude

Peter Yang demonstrated how a few prompts let Claude call HeyGen to generate multiple virtual characters and record talking clips, assembling them into a creator video wall. According to Jake Moran, he didn't even ask to be put in the frame—Claude decided on its own to add his avatar in the middle position. The full tutorial is included in a 40-minute AI video editing program.

Peter Yang shares method for generating a creator video wall with Claude

Peter Yang demonstrated how a few prompts let Claude call HeyGen to generate multiple virtual characters and record talking clips, assembling them into a creator video wall. According to Jake Moran, he didn't even ask to be put in the frame—Claude decided on its own to add his avatar in the middle position. The full tutorial is included in a 40-minute AI video editing program.

View story:Peter Yang shares the method for generating a creator video wall with Claude

SuperCLUE Releases Overall Ranking of I2V Image-to-Video Models

On October 23, 2026, SuperCLUE updated and released the overall ranking of I2V image-to-video models. According to reports by 21ic Electronics Network and Phoenix Technology, MiniMax H3 from MiniMax ranked first in China with an overall score of 71.46, and its application capability score of 75.82 was the highest on the entire list; Alibaba's Wan 3.0 ranked second with 66.63 points, with its basic capability score of 68.51 overtaking MiniMax H3; ByteDance's Seedance 2.5 and Kuaishou's Kling V3 Pro tied for third with 64.63 points and 64.06 points, respectively. The reports stated that overseas models were included only for comparison and were not counted in the official ranking, among which Gemini Omni 1.1 Flash scored 67.38 points and Grok Imagine Video 1.5 scored 65.79 points. At present, there is no further progress following the release of this ranking.

fal claims FLUX 3 Max can generate a 10-second video in just 12.45 seconds

fal claims that FLUX 3 Max, a post-trained video model it released in collaboration with bfl_ai, takes 12.45 seconds to generate a 10-second video. The same source says that, by comparison, Seedance 2.5 takes 277.8 seconds to generate the same video; the two share an Elo score of 1,139 on a third-party benchmark evaluation. The model is tied for first place in an independent third-party benchmark's image-to-video quality ranking, and can generate 10-second 720/768p videos in about 12 seconds. fal also says that FLUX 3 Max will drop in price by 75% over the next three weeks. Currently this news has only been released by fal, with no subsequent independent developments yet.

fal claims FLUX 3 Max generates a 10-second video in just 12.45 seconds

fal claims that FLUX 3 Max, a post-trained video model it released in collaboration with bfl_ai, takes 12.45 seconds to generate a 10-second video. The same source says that, by comparison, Seedance 2.5 takes 277.8 seconds to generate the same video; the two share an Elo score of 1,139 on a third-party benchmark evaluation. The model is tied for first place in an independent third-party benchmark's image-to-video quality ranking, and can generate 10-second 720/768p videos in about 12 seconds. fal also says that FLUX 3 Max will drop in price by 75% over the next three weeks. Currently this news has only been released by fal, with no subsequent independent developments yet.

View story:fal claims FLUX 3 Max can generate a 10-second video in just 12.45 seconds

Raschka shares an eight-part video course on implementing reasoning models from scratch

English text is being prepared.

View story:English text is being prepared.

Tencent Hunyuan launches open-source interactive video world model Hy-Nomas

Tencent Hunyuan launched the open-source interactive video world model Hy-Nomas on October 24, 2026. According to official release information, the model supports interaction using text prompts and camera trajectories, and has long-horizon generation capability with world memory; it can be adapted to specialized domains such as robotics via lightweight LoRA, and has been optimized for inference on consumer-grade GPUs. The official release also provides website, GitHub, Hugging Face, and demo entry points. There are currently no subsequent development facts.

Tencent Hunyuan launches open-source interactive video world model Hy-Nomas

Tencent Hunyuan launched the open-source interactive video world model Hy-Nomas on October 24, 2026. According to official release information, the model supports interaction using text prompts and camera trajectories, and has long-horizon generation capability with world memory; it can be adapted to specialized domains such as robotics via lightweight LoRA, and has been optimized for inference on consumer-grade GPUs. The official release also provides website, GitHub, Hugging Face, and demo entry points. There are currently no subsequent development facts.

View story:Tencent Hunyuan launches open-source interactive video world model Hy-Nomas