AI Video News
A daily news feed on AI video and AI filmmaking: new models and software, new features, platform rules, and the industry moves that change what’s worth using. Newest first, with the source and check date on every entry.
Higgsfield AI Adds Genjutsu Tool for Object and Character Swaps
Higgsfield AI launched Genjutsu, a video editing tool that replaces objects, locations, or characters in a video while keeping other elements unchanged. The tool is part of Higgsfield’s existing plans.
Source: higgsfield.ai · Second source: krea.ai · Checked October 16, 2026
Kling 4.0 Is Now Available to All Users
Kuaishou has opened the full Kling 4.0 model to all users, after Kling 4.0 Flash went to a limited early-access group on September 28. Kling 4.0 generates clips of up to 30 seconds in one pass and takes up to 10 keyframes and up to 15 mixed references (Omni Reference). Kuaishou also lists steadier large-scale motion and complex camera moves, more accurate lip sync, video extension, more languages and accents, and a redesigned creation page. 10-bit HDR output at 1080p and 4K is announced but still marked as coming soon.
On GraiLogic: We have already ranked Kling 4.0. Until independent tests and public leaderboard results arrive, its score is a clearly marked provisional projection, based on how much previous major model generations gained on public blind-vote leaderboards. It replaces the measured Kling 3.0 score in our text-to-video and image-to-video rankings, and switches to measured data as soon as it exists. Read the review →
Source: x.com/Kling_ai · Second source: runtimewire.com · Checked October 16, 2026
ElevenLabs CLI v1.1.0 Adds Request Attribution for Agents
ElevenLabs released CLI v1.1.0 with command-level request attribution for ElevenAgents workflows, following v1.0.0 the prior week. The update helps developers and video creators track usage across voice agent pipelines.
On GraiLogic: ElevenLabs offers a free plan with 10k credits, with paid plans starting at $6 per month for commercial use. Read the review →
Source: elevenlabs.io · Second source: elevenlabs.io · Checked October 15, 2026
ElevenLabs Music Dubbing SDK Gains API Updates
ElevenLabs updated its ElevenAgents Music Dubbing SDK and Dubbing v2 API, which translates audio and video into over 90 languages via API access. This gives developers more programmatic control over dubbing workflows for video localization.
On GraiLogic: ElevenLabs plans start at $6 per month with a free 10k-credit tier, though commercial use requires a paid subscription. Read the review →
Source: elevenlabs.io · Second source: elevenlabs.io · Checked October 15, 2026
Pippit AI Adds SeedAudio 2.0 for Text-to-Audio Generation
Pippit AI integrated SeedAudio 2.0, enabling full AI audio generation from text prompts directly within its platform. Video creators can now generate voiceovers and sound directly inside Pippit’s editing workflow.
On GraiLogic: Pippit AI plans start at $15 per month with a free tier offering 150 credits weekly. Read the review →
Source: pippit.ai · Second source: gettysburgtimes.com · Checked October 15, 2026
Kuaishou Announces Kling 4.0 for October 2026 Launch
Kuaishou revealed Kling 4.0 will launch in October 2026 with 4K resolution, stereo audio, improved lip sync, 10-bit HDR, and native generation up to 30 seconds.
Source: theresanaiforthat.com · One source, not independently verified · Seen October 14, 2026
YouTube Adds Automatic Labels for AI-Generated Content
YouTube now automatically applies visible labels to AI-generated videos, shown below the video and as overlays on Shorts. This changes disclosure requirements for creators uploading AI-made content to the platform.
Source: youtube.com · One source, not independently verified · Seen October 12, 2026
TikTok Rolls Out Agentic AI Shopping Tools for Advertisers
TikTok introduced agentic AI advertising tools covering discovery through conversion, including a Shopping Assistant and Buy Direct features. This changes how creators and brands can sell directly through TikTok content.
Source: tiktok.com · One source, not independently verified · Seen October 11, 2026
Seedance 2.5 Arrives on Pippit AI With Precise Segment Editing
Pippit AI has added Seedance 2.5, which allows precise segment-level editing within generated videos. The update gives video creators finer control over edits without regenerating entire clips.
On GraiLogic: Pippit AI offers a free plan with daily credits, while commercial use requires a paid tier starting around USD101 per year. Read the review →
Source: pippit.ai · Second source: finance.yahoo.com · Checked October 10, 2026
HeyGen Open-Sources Real-Time Avatar Stack With GPT-Live-1
HeyGen released an open-source real-time avatar stack built on GPT-Live-1, launched a Code2Video benchmark, and added Professional Voice Clone to its API. The API access to professional voice cloning lets developers integrate cloned voices directly into video pipelines.
On GraiLogic: HeyGen’s Creator plan starts at $29 per month, or $24 per month billed annually, with a free plan available. Read the review →
Source: heygen.com · One source, not independently verified · Seen October 8, 2026
Google Sets Gemini Omni Flash as Default Video Model
Google has made Gemini Omni Flash the default video generation model for video creators using its tools. This changes which model users get by default when generating video through Google’s platform.
Source: ai.google.dev · One source, not independently verified · Seen October 7, 2026
ElevenLabs Launches Flows Node-Based Canvas for Media Models
ElevenLabs Flows is a new node-based canvas supporting over 35 image and video models, with a Flow Agent for automating generation workflows. Video creators get a single interface to chain multiple models together.
On GraiLogic: ElevenLabs plans start at $6 per month for the Starter tier with 30000 credits, and a free plan with 10000 credits is also available. Read the review →
Source: eleven.labs · One source, not independently verified · Seen October 6, 2026
Viggle AI Cuts Annual Prices for Pro, Live, and Max Plans
Viggle AI’s annual billing now prices Pro at $7.99/month, Live at $15.99/month, and Max at $63.99/month, down from $9.99, $19.99, and $79.99 respectively.
On GraiLogic: Viggle’s free plan still offers 5 videos per day, but commercial use requires one of the paid tiers. Read the review →
Source: viggle.ai · Second source: aimiracle.ai · Checked October 5, 2026
Lightricks Releases Open-Source LTX-2.3 With Native 4K and Audio
Lightricks launched LTX-2.3 as an open-weight model line offering native 4K output and audio generation, giving video creators a downloadable alternative to closed commercial models.
Source: ltx.io · One source, not independently verified · Seen October 5, 2026
ByteDance Seedance 2.0 Adds Multi-Shot Control and Native Audio
ByteDance released Seedance 2.0 with multi-shot control and native audio generation, and the model reached the top spot on the Artificial Analysis video leaderboard.
Source: doubao.com · One source, not independently verified · Seen October 5, 2026
InVideo Adds Sora 2 and Veo 3.1 Generation Inside Its Editor
InVideo integrated direct access to Sora 2 and Veo 3.1 generation models within its editing interface, letting users generate clips without leaving the platform.
On GraiLogic: InVideo’s Starter plan runs $20 per month with no free tier currently listed. Read the review →
Source: invideo.io · One source, not independently verified · Seen October 5, 2026
Google Launches Budget Veo 3.1 Lite With Native Audio
Google released Veo 3.1 Lite, a lower-cost tier priced at $0.05 per second that includes native audio generation, widening access to Veo-quality video generation.
Source: deepmind.google · One source, not independently verified · Seen October 5, 2026
Descript’s Underlord Co-Editor Gains Auto-Cutting and B-Roll Tools
Descript enhanced its Underlord AI co-editor to automatically tighten cuts, insert B-roll, and reformat content for different platforms.
On GraiLogic: Descript’s plans start at $16/month for Hobbyist and $24/month for Creator, with annual billing offering bonus credits. Read the review →
Source: descript.com · One source, not independently verified · Seen October 5, 2026
Synthesia Adds Video Translator and Interactive Avatars Beta
Synthesia has added a Video Translator feature and launched Interactive Avatars in beta, expanding its avatar video platform for creators needing localization and interactive formats.
On GraiLogic: Synthesia is listed on our site starting at $29/mo for the Starter plan, with a free Basic tier offering 10 minutes per month with watermark. Read the review →
Source: synthesia.io · Second source: eesel.ai · Checked October 3, 2026
Meta Launches Muse Realtime Avatar From a Single Photo
Meta has launched Muse, a realtime avatar tool that converts a single photo into a live speaking video in under a second. This could offer a new option for creators needing fast talking-head content.
Source: meta.com · One source, not independently verified · Seen October 2, 2026
Alibaba Wan 3.0 Tops Artificial Analysis Video Leaderboard
Alibaba’s Wan 3.0 video model has ranked first on the Artificial Analysis leaderboard, with native audio generation built into its outputs. This marks a new top-tier model generation for video creators comparing text-to-video options.
Source: analyticsinsight.net · One source, not independently verified · Seen October 1, 2026
Colossyan Adds Cora AI Agent to Generate Video Drafts From Docs
Colossyan introduced Cora, an AI agent that turns documents, PDFs, PowerPoint files, and scripts into draft video plans for user approval. This lets creators start a video project directly from existing written material instead of building scenes from scratch.
Source: colossyan.com · Second source: colossyan.com · Checked September 26, 2026
Adobe Premiere Pro Improves Firefly-Powered Editing Tools
Adobe updated Premiere Pro’s Firefly integration, improving accuracy for Generative Extend, Content-Aware Fill, and Auto Reframe. These tools help editors extend clips, remove unwanted objects, and reframe footage for different aspect ratios directly inside the timeline.
Source: pixflow.net · One source, not independently verified · Seen September 26, 2026
Suno Launches Studio 2.0 Browser-Based DAW for Premier Users
Studio 2.0 is a browser-based digital audio workstation with MIDI support, a wavetable synthesizer, and chat-based editing, available to Premier subscribers as of August 13, 2026. This gives video creators more control over composing and editing music tracks before exporting them into a project.
Source: suno.com · Second source: musicbusinessworldwide.com · Checked September 25, 2026
Captions.ai Releases Mirage Avatar X Model
Captions.ai has released Mirage Avatar X, described as its most advanced avatar model, giving video creators a new option for AI-generated talking avatars.
Source: captions.ai · Second source: mirage.app · Checked September 25, 2026
PixVerse Offers Limited-Time Discount on MiniMax H3 With 2K Modes
PixVerse added MiniMax H3 with 2K resolution modes and is running a 14-day promotion starting August 20, cutting Pro, Premium, and Team Premium plans by 50 percent and Ultra and Team Ultra plans by 60 percent. Commercial use still requires a paid plan.
Source: pixverse.ai · Second source: x.com · Checked September 24, 2026
Google DeepMind Releases Veo 3.1
Veo 3.1 is now available from Google DeepMind. Pricing remains unchanged at $19.99/mo for Google AI Pro, $0.05/sec via the Lite API, with a free tier offering 100 credits per month.
Source: deepmind.google · Second source: buildfastwithai.com · Checked September 23, 2026
Higgsfield Adds 3D Jutsu Workspace for AI-Native 3D Scenes
3D Jutsu is a new AI-native 3D workspace on Higgsfield, powered by Supercomputer, that builds editable 3D scenes from a prompt or reference image and turns them into video.
Source: higgsfield.ai · Second source: higgsfield.ai · Checked September 23, 2026
Runway Unveils Solaris Interface World Model
Runway introduced Solaris, an interface world model that generates interactive user interfaces directly instead of through coding. Commercial use of Runway output still requires a paid plan.
Source: runwayml.com · Second source: runway.com · Checked September 23, 2026
Runway Adds GWM Worlds 2 Real-Time Simulation Model
GWM Worlds 2 on Runway generates real-time interactive simulations at 720p and 24 fps with 48 kHz audio, giving creators a new option for interactive world generation.
Source: runwayml.com · Second source: runway.com · Checked September 22, 2026
MiniMax H3 Max Arrives on Higgsfield With New Plan Pricing
Higgsfield now offers the MiniMax H3 Max model for text-to-video and image-to-video generation, alongside updated annual plan pricing: Starter at $19/mo for 270 credits, Plus at $47/mo for 1200 credits, and Ultra at $99/mo for 3000 credits.
Source: higgsfield.ai · Second source: creatify.ai · Checked September 22, 2026
Luma Integrates Seedance 2.5 Model Into Dream Machine
Luma has added Seedance 2.5 to its platform, giving users access to the new model within Dream Machine and Agents. This expands the range of generation models available without leaving Luma’s interface.
Source: lumalabs.ai · Second source: lumalabs.ai · Checked September 21, 2026
OpenAI Sora API to Shut Down September 24, 2026
OpenAI will end developer access to the Sora API on September 24, 2026, cutting off programmatic video generation through this route. Developers relying on the API for video pipelines will need to migrate to alternative tools before the shutdown date.
Source: en.wikipedia.org · One source, not independently verified · Seen September 21, 2026
Luma AI Launches Native Arabic Interface
Luma AI has introduced an Arabic-language interface aimed at Saudi and Arabic-speaking creators, making the platform more accessible to that region. This does not change pricing or plans for existing users.
Source: lumalabs.ai · Second source: lumalabs.ai · Checked September 20, 2026
Hailuo AI Video Cuts Free Tier to Daily Generation Limits
Hailuo AI Video has reportedly reduced its free tier from unlimited generations to around 3 to 5 per day since its initial launch. Video creators relying on Hailuo’s free access should check current limits before planning workflows.
Source: heygen.com · One source, not independently verified · Seen September 20, 2026
Luma Adds Scenes Feature to Dream Machine
Luma has released a Scenes feature within its platform, expanding tools available to video creators using Dream Machine.
Source: lumalabs.ai · Second source: itbrief.co.uk · Checked September 19, 2026
DeeVid AI Integrates MiniMax H3 With Native 2K, 15-Second Clips
DeeVid AI now offers the MiniMax H3 model, which generates 2K video up to 15 seconds long with synchronized stereo audio from text, image, video, or audio inputs.
Source: deevid.ai · Second source: minimax.io · Checked September 19, 2026
AI Content Labeling Rules Take Effect Under EU AI Act and SB 942
Mandatory labeling requirements for AI-generated content under the EU AI Act Article 50 and California SB 942 became enforceable on August 2, 2026. Video creators distributing AI-generated content in these jurisdictions must now comply with disclosure rules.
Source: ec.europa.eu · One source, not independently verified · Seen September 19, 2026
Kling AI Adds Video 3.0 Motion Control Update
Kling AI has rolled out a major update to Video 3.0 called Motion Control, expanding its motion direction capabilities for generated clips. The update is confirmed for rollout in early 2026.
Source: kling.ai · Second source: en.ai-creators.tech · Checked September 18, 2026
Tavus Launches Deployments for Sending PALs Live
Tavus introduced Deployments, a new feature that lets creators push their PAL conversational agents into live production environments. This adds a distribution layer on top of Tavus’s existing conversational video tools.
Source: tavus.io · Second source: tavus.io · Checked September 18, 2026
Kling AI Adds Native 4K and Series Mode to Image 3.0 Omni
Kling AI’s Image 3.0 Omni now supports native 2K and 4K image output along with a new Series Mode for generating consistent image sequences. This expands the tool’s image generation capabilities for video creators building visual assets.
Source: kling.ai · Second source: kling.ai · Checked September 18, 2026
Higgsfield Aggregates Veo, Kling, Sora, Wan, and Seedance Models
Higgsfield is a multi-model platform that gives creators unified access to Veo, Kling, Sora, Wan, and Seedance video models, along with shared editing and consistency controls. This lets video creators compare and use multiple leading models from a single interface.
Source: higgsfield.ai · One source, not independently verified · Seen September 18, 2026
Tavus Releases Phoenix-4.5 Real-Time Human Rendering Model
Tavus launched Phoenix-4.5, a real-time human rendering model the company describes as its fastest and most expressive to date. The model targets developers building real-time conversational video avatars.
On GraiLogic: Tavus plans start from 59 dollars per month for the Starter tier with a free tier offering 25 conversational minutes. Read the review →
Source: tavus.io · Second source: tavus.io · Checked September 17, 2026
Descript Adds Claude Fable 5 to Underlord Model Picker
Descript now lets users select the Claude Fable 5 model within the Underlord panel for AI-assisted editing tasks. The model is available as an option alongside existing choices in the model selector.
On GraiLogic: Descript’s Creator plan starts at $24 per month and requires a paid plan for commercial use. Read the review →
Source: descript.com · Second source: descript.canny.io · Checked September 16, 2026
ElevenLabs Adds Conversation Summaries to ElevenAgents
ElevenAgents now supports lightweight conversation summaries with title, transcript summary, and call outcome fields, plus an optional max_messages parameter and new pagination. This helps creators managing voice agent conversations review interactions more efficiently.
On GraiLogic: ElevenLabs offers a free plan with 10k credits, with paid plans starting at $6 per month for 30k credits. Read the review →
Source: elevenlabs.io · Second source: elevenlabs.io · Checked September 16, 2026
ElevenLabs Launches Flows API for Async Video and Speech Generation
ElevenLabs released ElevenCreative Flows API, an asynchronous system for generating video, images, and speech with create, list, and retrieve endpoints. This gives developers a way to build automated content pipelines using ElevenLabs generation capabilities.
On GraiLogic: ElevenLabs plans start at $6 per month with a free tier offering 10k credits. Read the review →
Source: elevenlabs.io · Second source: elevenlabs.io · Checked September 16, 2026
Adobe Firefly Expands With Partner Video Models and Credits System
Adobe Firefly now offers access to partner video generation models alongside a generative credits system. This broadens the range of video generators available to creators working within Adobe’s ecosystem.
Source: zapier.com · One source, not independently verified · Seen September 16, 2026
Cartesia Adds Keyterm Prompting and Turn Detection to Ink-2
Cartesia’s Ink-2 model now supports keyterm prompting and configurable turn detection thresholds, including turn-start, turn-end, and eager-end settings. This gives creators more control over how narration or voice agents handle pacing and pronunciation.
On GraiLogic: Cartesia offers a free tier with 20K credits per month, alongside paid plans starting at 5 dollars per month for 100K credits. Read the review →
Source: cartesia.ai · Second source: docs.cartesia.ai · Checked September 13, 2026
Descript Adds AI Image Editing With Drawn Annotations
Descript now lets users draw directly on an image and describe the desired change to guide AI-based edits. This adds a visual editing option alongside its existing text-based tools.
On GraiLogic: Descript includes a free plan, with paid tiers starting at 16 dollars per month for the Hobbyist plan. Read the review →
Source: descript.com · Second source: help.descript.com · Checked September 13, 2026
Tavus Launches Sparrow-2 Real-Time Conversation Model
Sparrow-2 handles turn-taking, interruptions, and acoustic scenes as a unified system for real-time conversational video. It is aimed at developers building interactive avatar and voice experiences.
On GraiLogic: Tavus offers a free tier with 25 conversational minutes and paid plans starting at 59 dollars per month for 100 CVI minutes. Read the review →
Source: tavus.io · Second source: tavus.io · Checked September 12, 2026
InVideo Launches Enterprise Tier With Isolated Environments
InVideo for Enterprise adds dedicated isolated environments, full audit logging, access controls, and commercial usage indemnification for organizations. It targets teams needing tighter governance over AI video generation.
Source: invideo.io · Second source: facebook.com · Checked September 12, 2026
Adobe Adds Generative Media Tool to Premiere Pro
Adobe introduced a Generative Media Tool inside Premiere Pro that lets editors generate video and audio directly within their projects. The feature integrates AI generation into the standard editing workflow rather than requiring a separate tool.
Source: kingy.ai · One source, not independently verified · Seen September 12, 2026
InVideo Agent One Tops Physion-Arc 1.0 Benchmark
InVideo Agent One ranked first on the independent Physion-Arc 1.0 benchmark, confirmed by a re-evaluation dated late July 2026. The result signals strength for InVideo’s agentic video workflows relative to competitors.
Source: invideo.io · Second source: manilatimes.net · Checked September 11, 2026
InVideo Launches Agent Two With Workflows and Enterprise Tools
InVideo released Agent Two, adding Workflows, Enterprise, Folders, Connect, Stock, Slate, Notebooks, Playbooks, Multi-Input Intelligence, and Expert Agents. The update expands InVideo beyond single-video generation toward broader production and team workflows.
Source: invideo.io · Second source: creatify.ai · Checked September 11, 2026
Magic Hour Adds Kling 2.6 Model Support
Magic Hour now offers Kling 2.6 for text-to-video and image-to-video generation, giving users an additional model choice within the platform.
Source: docs.magichour.ai · Second source: overchat.ai · Checked September 11, 2026
HeyGen Adds Video Podcast, HyperFrames, 30-Min Avatar Videos
HeyGen introduced Video Podcast creation, HyperFrames, 30-minute avatar videos, and website-to-video tools. The update expands what creators can produce within existing plans.
On GraiLogic: HeyGen remains listed with a free plan and paid tiers starting at 29 dollars per month for commercial use. Read the review →
Source: heygen.com · Second source: flowith.io · Checked September 10, 2026
EU AI Act Requires Disclosure of Synthetic Media of Real People
As of August 2026, EU enforcement makes it legally mandatory to disclose AI-generated synthetic media depicting real people. Video creators distributing content in the EU must add clear disclosure to comply.
Source: ec.europa.eu · One source, not independently verified · Seen September 10, 2026
YouTube Enforces Auto-Labeling and Disclosure for AI Content
YouTube now automatically labels AI-generated videos and requires creators to disclose synthetic media depicting real people. This affects how video creators using AI tools must tag and publish their uploads.
Source: support.google.com · One source, not independently verified · Seen September 10, 2026
Suno Sets Download Limits Starting September 2026
Suno’s updated download policy takes effect September 3, 2026, capping free downloads at 7 lifetime tracks, Pro at 20 per month, and Premier at 60 per month. The change affects how much finished audio creators can pull from the platform for use in their videos.
Source: suno.com · Second source: musicbusinessworldwide.com · Checked September 9, 2026
Google Veo 3.1 Adds Single-Pass Synchronized Audio
Veo 3.1 now generates video with synchronized audio in one pass, including 48kHz dialogue support, reducing the need for separate audio syncing steps.
On GraiLogic: Google Veo is listed at 19.99 dollars per month for the AI Pro plan, with API pricing at 0.05 dollars per second and a free tier offering 100 credits monthly. Read the review →
Source: deepmind.google · One source, not independently verified · Seen September 9, 2026
Prices and plan names change often, and vendors do not always announce it. If you spot something here that no longer matches the vendor’s own page, tell us at editor@grailogic.com and we will re-check it. See also our Methodology and Affiliate Disclosure.