更新
AI 视频模型新闻与版本归档
External release links, version notes, 和 company milestones are rendered into static HTML so 有来源支持 updates are visible without JavaScript.
最新来源链接
Sorted by public date when available. 部分支持 or future-looking claims stay visible as such in profile pages.
- Seedance is making a fresh push into education，并study app Gauth 说明 ByteDance 的分发策略，以及报道中的 Seedance 2.5 教育场景。
- ByteDance's big bet on AI Useful company-level 背景 for ByteDance's AI investment 和 Volcano Engine commercialization.
- Elon Musk's xAI sues Minnesota over its first-in-the-nation law banning 'nudification' technology Safety 和 regulatory 背景 relevant to consumer trust scoring.
- xAI's last-minute scramble to stop Minnesota's anti-nudification app law Adds policy-risk 背景 around misuse / generative image/video tools.
- PixVerse API changelog: V6 Reference-to-Video omni mode V6 reference-to-video adds reference_mode omni 和 video_参考素材.
- PixVerse closed a Series C 延展 和 announced expansion into interactive entertainment. 公司 milestone source.
- Luma — Official Information for AI Assistants | Luma Official structured source for current Luma model naming 和 Ray version.
- Runway API changelog: Veo negative prompt support Runway API adds optional negativePrompt for veo3, veo3.1, 和 veo3.1_fast video requests.
- 模型 cards DeepMind 官方模型卡索引提供 Veo 安全与版本更新。
- Sora 2 模型 | OpenAI API Official API model page for Sora 2 pricing 和 endpoints.
- What to know about the Sora discontinuation OpenAI Help Center says Sora web/app was discontinued on 2026-04-26 和 the API is scheduled for 2026-09-24.
- 模型 Official Luma Agents API model capability 和 pricing surface.
- Pika Official FAQ for current Pika model, durations, resolution, API access, 和 credit costs.
- LTX GitHub news Official repository README news for LTX-Video 和 LTX-2. README fetched, but no dated news bullet was parsed.
- Video Generation | SpaceXAI Docs Official xAI video generation API capability docs.
- Krea 2 Community License Agreement was published for downloadable Krea 2 model variants. 公司 milestone source.
- Grok Imagine Video 1.5 Primary official release for Grok Imagine Video 1.5.
- Generate videos using text prompts Current official product instructions 和 model-selection details.
- Luma Introduces Ray3.2 模型 & API: Complete Creative Control for Video Generation Primary release for Ray3.2 和 its API.
- xAI announced Grok Imagine 1.5 preview. 公司 milestone source.
- Introducing Runway MCP Important signal that Runway is moving into agent-connected creative 工作流.
- ShengShu unveiled Vidu Claw, an AI CMO product powered by Vidu Q3. 公司 milestone source.
- NVIDIA Cosmos release notes documented continued model updates through 2.2. 公司 milestone source.
- Higgsfield published a 2026 overview describing its multi-model creative platform. 公司 milestone source.
- Vidu API update: Q3 reference-to-video models Reference-to-video adds viduq3-mix, viduq3-turbo, 和 viduq3 models.
- Veo 3.1 Lite 模型 Card Veo 3.1 Lite 官方模型卡，可用于跟踪安全和能力信息。
- Deprecated Standalone Tools Official deprecation table for legacy Magic Tools 和 standalone video utilities.
- Introducing Real-Time Video Chat for Agents Official Pika 研究 post introduces PikaStream1.0 for real-time video chat agents.
- Create, edit 和 share videos at no cost in Google Vids Google announced Veo 3.1-powered video generation access in Google Vids，并monthly free generations for personal accounts.
- PixVerse Launches V6, Advancing AI Video Generation Across Creative 和 Agentic 工作流 Primary official V6 release post.
- SkyReels privacy policy identified Skywork AI Pte Ltd. as operator. 公司 milestone source.
- Seedance 2.0 Official Launch Primary official release for the 2.0 multimodal audio-video model.
- Seedance 2.0 Official product page describing capabilities 和 基准测试 framing.
- Is the LTX model open source? Useful official clarification on weights 和 self-hosting.
- Pika's terms identified Mellis, Inc. as the service operator 和 described text/image/video generation services. 公司 milestone source.
- Kling VIDEO 3.0 Omni 模型 User Guide Product guide describing multimodal inputs, 原生音频, 和 storyboarding.
- Kling AI 3.0 series fully launches globally Official corporate release for Kling 3.0 和 Kling 3.0 Omni.
- 29, 2026: 🎉 We launched the API for the SkyReels-V3 models on the apifree.ai. Official repository README news for SkyReels V-series releases.
- Veo 3.1 Ingredients to Video: More consistency, creativity 和 control Primary official update for Veo 3.1 features.
- Lightricks Open-来源 LTX-2 Primary open-weight release announcement.
- Runway Advances Video Generation 和 World 模型，并 NVIDIA Rubin Platform 公司 研究 infrastructure signal for future model development.
- Lightricks released full LTX-2 model weights. 公司 milestone source.
- Kling AI Video 和 Image Generator Live product page for current consumer tools 和 API navigation.
- V6 API capability map for production feature comparisons.
- Text to Video 模型 Third-party developer catalogue useful for 跟踪 supported PixVerse variants.
- Vidu Q3 AI Video 模型，并原生音频 Official Q3 consumer feature page.
- 模型 Map Official API model map for Q3 variants 和 capabilities.
- Artificial Analysis Text to Video Leaderboard Independent leaderboard reference for Vidu model variants.
- Veo Official 模型家族 page 和 capability overview.
- Generate videos，并Veo 开发者 access documentation for current video generation APIs.
- Sora Official landing page to recheck current availability before publication.
- Product 更新 & Changelog Operational release source for model access 和 feature changes.
- Ship cinema, not content. API capability page for multi-keyframe, V2V, HDR, 和 EXR comparisons.
- 新闻 Official release archive for Ray 和 Luma product updates.
- Pika Current official consumer product page.
- AI video model leaderboard Useful third-party reference for model coverage 和 consumer-facing comparison.
- LTX-2 Official code package for inference 和 LoRA training.
- grok-imagine-video-1.5 API model card 和 pricing source.
- Create videos，并the fast, simple, free AI 视频生成器 Consumer-facing feature page 和 partner model list.
- SkyReels createVideo Live product UI source for current 工作流 和 V4 Fast signal.
- SkyReels API Platform Commercial access source.
- CogVideo GitHub repository Current official source for checkpoints 和 usage.
- Open-source 图生视频 leaderboard Independent open-model 背景 for ranking.
- happyhorse-1.1-i2v Strongest official source currently found for HappyHorse access.
- Mochi GitHub repository Primary technical source for open-model users.
- Genmo 公司 和 product entry point.
- Replicate Text-to-Video collection Hosted inference 和 open-model discovery 背景.
- stable-video-diffusion-img2vid-xt Current model-card source for the XT checkpoint.
- NVIDIA Cosmos Official product/platform page.
- NVIDIA Cosmos GitHub 开发者 source for open resources.
- Video Generation 来源 to verify current model IDs.
- Wenxin Workshop Official Baidu AI Cloud entry point for model access 研究.
- MiracleVision Official product entry point for follow-up model mapping.
- Haiper Official consumer product page.
- Pricing Useful source for cost 和 model capability comparison.
- Higgsfield AI Official product entry point.
- AI Video Generator Consumer feature page.
- Generating Video on Leonardo.Ai Current user 工作流 source.
- AI Video Generator Consumer product page.
- Video to Video Current 工作流 documentation.