awesome-text-to-video

Awesome-Text-To-Video

A curated, visually-organized list of Text-to-Video (T2V) products, open-source models, research papers, datasets, and benchmarks.

---

Table of Contents

• What's New
• Commercial Text-to-Video Products
  • Generative Text-to-Video Models
  • AI Avatar / Presenter Tools
  • All-in-One Creative Platforms
  • Discontinued / Historical
  • Featured Product Cards
  • Related Tools
  • How to Choose
• Open-Source Models & Toolkits
  • Quick Start
• Research Papers
  • 2026
  • 2025
  • 2024
  • 2023
  • 2022
  • Before 2021
• Datasets & Benchmarks
• Contributing
• License
• References

---

What's New

2026 Update: The T2V landscape has shifted dramatically. OpenAI discontinued the consumer Sora app in early 2026, while open-source models (Wan 2.2, HunyuanVideo 1.5, LTX-2.3) and commercial alternatives (Runway Gen-4.5, Seedance 2.0, Kling 3.0, Veo 3.1) now dominate.

Data collected as of June 2026. Prices, features, and availability change quickly — always check the official website before making a decision.

Key trends:

• Native 4K and synchronized audio are becoming standard.
• Open-source models now rival closed commercial systems.
• Multi-shot storytelling and character consistency are the new battlegrounds.
• Inference acceleration (sliding tile attention, token carving, caching) makes local generation feasible.

---

Commercial Text-to-Video Products

Closed-Source Platforms

Generative Text-to-Video Models

Product	Maker	Best For	Max Output	Highlights	Link
🎬 Runway Gen-4 / Gen-4.5	RunwayML	Professional creators	1080p (4K upscale)	@reference consistency, world-class physics, Aleph editing	runwayml.com
🐉 Kling 3.0	Kuaishou	Motion-heavy / cinematic	1080p, 15 s	Best-in-class motion, generous free tier	klingai.com
🔍 Veo 3 / Veo 3.1	Google DeepMind	4K broadcast production	Native 4K	Scene extension, native audio + lip-sync	deepmind.google
🌱 Seedance 2.0	ByteDance	Multi-shot storytelling	2K, 60 s multi-shot	12 mixed inputs, native audio-video joint gen	seedance.tv
✨ Luma Dream Machine	Luma AI	Action / sports / physics	4K	Realistic motion blur, fluid dynamics	lumalabs.ai
🎭 Pika 2.5 / 3.0	Pika Labs	Social / stylized content	2K	Fast, cheap, strong style transfer	pika.art
🌀 Hailuo AI	MiniMax	Realistic humans / prompt adherence	1080p, 10 s	Strong physical realism, #1 in China	hailuoai.video
🌀 MiniMax H3 (Third-Party)	MiniMax3.org (independent)	Cinematic text/image/video/audio-reference generation	2K	768p/2K output and native audio; third-party platform, not an official MiniMax product	minimax3.org
🎬 PixVerse V6	PixVerse	Anime / stylized content	1080p, 15 s	Character consistency engine, 20+ camera controls, native audio	pixverse.ai
🎥 Vidu Q1/Q2	Shengshu / Tsinghua	Highly consistent T2V	1080p, 16 s	U-ViT backbone, subject consistency, 1080p generation	vidu.com
🔬 Lumiere	Google DeepMind	Research T2V / I2V / editing	720p	Space-time U-Net, single-pass temporal generation	lumiere-video.github.io
🎞️ Seele TV	Seele AI	Cinematic sequence workflows	Varies by selected model	Visual references, shot-level camera direction, and continuity-oriented video creation	seele.tv

AI Avatar / Presenter Tools

Product	Maker	Best For	Max Output	Highlights	Link
🧑‍💼 Synthesia	Synthesia	Corporate training / avatars	1080p	100+ avatars, 130+ languages	synthesia.io
🎤 DeepBrain AI	DeepBrain AI	Hyper-realistic avatars	1080p	PPT-to-video, chroma key, native AI anchors	deepbrain.io
🎥 HeyGen	HeyGen	AI avatars / translation	1080p	100+ avatars, multilingual voice clone, video translation	heygen.com
🏢 Colossyan	Colossyan	Corporate training / e-learning	1080p	SCORM export, branching quizzes, dual-avatar scenes	colossyan.com
⚡ Elai.io	Elai	Document-to-video automation	1080p	URL/PPTX-to-video, 100+ languages, ~1.5 min rendering	elai.io
🖼️ D-ID	D-ID	Talking photos / short outreach	1080p	Animate any photo, live streaming API, cheapest entry	d-id.com
📺 Hour One	Hour One	Enterprise presentations	1080p	Broadcast-ready virtual presenters, clean UI	hourone.ai
🧑‍🏫 Vidnoz AI	Vidnoz	Budget avatar videos / training	1080p	Cheap avatar generation, multilingual voiceovers	vidnoz.com
🎨 Steve.AI	Steve.AI	Animated explainer videos	4K	Multiple animation styles, character library	steve.ai

All-in-One Creative Platforms

Product	Maker	Best For	Max Output	Highlights	Link
🛠️ Krea AI	Krea	Model-agnostic access	Varies	Unified UI for Kling, Hailuo, Luma, Runway, Pika	krea.ai
🌀 Morphic / Morph Studio	Morph Studio	All-in-one creative platform	Varies	15+ models, canvas editing, custom model training	morphic.com
✂️ InVideo AI	InVideo	Social media / marketing	1080p	5000+ templates, AI script-to-video workflow	invideo.io
🎬 YumCut	YumCut	Self-hosted vertical shorts	9:16	Script, voice, visuals, captions, and automation API	yumcut.com · GitHub
🎩 Magic Hour	Magic Hour	Multi-format creative suite	1080p	Face swap, talking photos, headshots, clothes swapper	magichour.ai
🛒 Creatify	Creatify	UGC-style ad generation	1080p	E-commerce focused, ad performance tracking	creatify.ai
🎞️ Vivideo	Vivideo	Model-agnostic short-video creation	Varies	Unified access to multiple T2V/I2V models, synced audio, text- and image-to-video	vivideo.ai
🎞️ cv.cm/v (Cloud Clipboard AI Studio)	Cloud Clipboard	Seedance-based video workflows	Varies	Queue-free Seedance 2.0, image generation, canvas, and short-drama agent	cv.cm/v
🎞️ Pixo	Pixo	Story-driven / storyboard-to-film	Varies	Story idea → storyboard → scene-by-scene generation → finished video, agent-assisted editing	pixo.video
📣 Advibly	Advibly	On-brand ad creative (UGC / image / video)	Varies	Brand + product context drives generation; UGC and talking-actor ads, carousels, voiceover and music; agent-native via MCP	advibly.com
📝 videos.social	videos.social	Content marketers / faceless drafts	Varies	Editable draft from blogs, PDFs, and prompts (script, scenes, voiceover); 1 free render; 1 credit = 1 render	videos.social
📅 Unsora	Shipped Studio	Generating and scheduling social media content	Varies	Video via Kling, Veo 3.1, and Seedance 2.0, plus images, talking avatars, music, and voiceovers; schedules posts to YouTube, TikTok, Instagram, and other platforms; REST API, MCP server, and CLI	tryunsora.com · GitHub
📱 MakeAIVideo	MakeAIVideo	Short-form social video (TikTok / Reels / Shorts)	Varies	Prompt, script or photo to finished video: script, AI or stock scenes, AI voiceover in 24 languages, captions and music; posts to social platforms; agent-native via MCP	makeaivideo.ai · MCP server
🎭 Katechon Arc	Katechon	AI short dramas / video ads	Varies	Idea, script, novel, or product brief → characters, storyboards, video, and voice in one agent-guided studio; remote MCP server	arc.katechon.technology
🎯 aiMakeVideo	aiMakeVideo	One-photo short-form video	Varies	Text/image-to-video plus image tools in one studio; 10 trend templates turn one photo into a 15 s 9:16 clip without a prompt; 9 UI languages	aimakevideo.com
🛍️ Mutator	Bitnook	Product videos for TikTok and Instagram	Varies	Starts from a product photo, a website or an app screenshot; short videos, photo slideshows and product photos; schedules to your own TikTok and Instagram; MCP server for agents	mutator.app
Sora2 Hub (Third-Party)	Sora2 Hub (independent platform)	Multi-model video and image generation	Varies	Text-to-video and image-to-video with a shared credit balance; third-party platform, not an official OpenAI product	sora2hub.org
GrokVideo Studio (Third-Party)	Grok Video Generator (independent)	Text-to-video and image-to-video	Varies	Uses the Grok Imagine API; 20 trial credits for new accounts; independent platform, not affiliated with or endorsed by xAI or X	grokvideogenerator.dev
PromptImage	PromptImage	AI image and video creation	Varies	Multi-model image generation, text-to-video workspace, and a prompt library	promptimage.cloud
🕺 Pixwit	Pixwit	Photo-to-video / meme dance effects	Varies	Freemium AI photo-to-video and meme dance effects; templates like Tiptoeing in My Jordans and Dime Dime Baby	pixwit.ai
🥋 PixelDojo	PixelDojo	AI image and video studio / AI director for ads and short films	Varies	80+ image and video models, plus Sensei, an AI director that makes full ads and short films; freemium	pixeldojo.ai
📢 pitchboom.io	pitchboom	Motion-design video ads from a product URL or brief	1080p (4K on eligible plans), up to 120 s	Script, editable storyboard, voice-over in 34 languages and a generated music bed; 16:9, 9:16, 1:1, 4:5 and 3:4, each recomposed rather than cropped; Free credits at sign-up, then paid plans from €30/month	pitchboom.io
🧬 RefVid	RefVid	Consistent product / character video	1080p, 30 s	Save a product or character once from 1–10 photos, then generate clips on Kling 4, Seedance 2.5 or Kling 3.0 Omni that keep the same label, face and outfit; up to 3 subjects per shot; free credits on sign-up	refvid.com
H3 MiniMax (Third-Party)	H3 MiniMax (independent)	Text-to-video and image-to-video with the MiniMax H3 model	768P or 2K, 4–15 s	Independent web app for the MiniMax H3 video model: text-to-video and image-to-video clips of 4–15 seconds at 768P or 2K; free trial credits; not an official MiniMax product	h3minimax.com
🎞️ Ideoy	Ideoy	Photo-to-video meme templates	Varies by template	Turns a photo into short vertical meme videos using preset trend templates	ideoy.ai
Ozil Goal AI (Third-Party)	Ozil Goal AI (independent)	Photo-to-video football template	14 s, vertical	Turns one photo into a goal-inspired clip; independent fan tool, not affiliated with Mesut Özil or any club, league, or competition	ozilgoalaivideo.org
🧰 NanoMaker AI Video	NanoMaker	Browser text-to-video and image-to-video	Varies	Multi-model video studio (incl. Seedance); start from a prompt or photo under one subscription	nanomaker.im/video
🦗 Cicadas.ai	Cicadas.ai	Shorts/UGC multi-model compare	Varies	One credit balance across Kling, Veo, Seedance, Wan, Hailuo; price shown before each generate	cicadas.ai
🎬 Seedance Now (Third-Party)	Seedance Now (independent)	Multi-model text-to-video and image-to-video	1080p, up to 30 s	Seedance 2.5, Kling 4.0 and Wan 3.0 in one browser studio; no watermark; Free signup credits subject to current offer; independent platform, not affiliated with ByteDance	seedancenow.com
🎥 Cyanvid	Cyanvid	Creators / marketers	Varies	Text/image-to-video with clear credit quotes before generate; multi-model studio	cyanvid.com

Discontinued / Historical

Product	Maker	Best For	Max Output	Highlights	Link
⚠️ Sora / Sora 2	OpenAI	Photorealism	1080p–8K	Consumer app shut down March 2026	openai.com/sora
⚠️ Haiper	Haiper	Fast creator clips	1080p	Consumer product shut down in 2026	haiper.ai

Featured Product Cards

Related Tools

Tool	What it does	Link
Podframes	Generates two-host AI podcast videos with scripted dialogue, voices, lip-sync, and captions.	github.com/Jellypod-Inc/podframes
TubePrompter	Converts existing videos into optimized text-to-video prompts for Sora, Veo, Runway, etc.	tubeprompter.com
Vadoo AI	AI shorts automation platform for faceless channels and social clips.	vadoo.tv
Omni-Rewriter	Open agentic prompt-expansion harness for image/video model dialects (schema + validation + bounded repair; expand ≠ generate).	github.com/WayneJin0918/Omni-Rewriter
BeatDesign	Open-source, local-first AI media workbench combining a Canvas, short-form video editor, shared Assets, and MCP control for image and video workflows.	github.com/BeatAPI/BeatDesign
minimax-h3-1000-prompts	Curated index of the MiniMax H3 1K prompt dataset: 3-field prompt anatomy, 10 hand-picked reusable prompts, and a model comparison. Interactive atlas of all 1,000 clips.	github.com/yangzhou-chaofan/minimax-h3-1000-prompts · neta.art atlas
VideoCreate	Splits a reference clip into timestamped shots with a per-shot sheet (subject, camera, motion, light, style, sound) and writes prompts adapted for Seedance, Kling 3.0, Wan 2.6 and Grok Imagine; includes a library of prompts tested on real generations.	videocreate.org
PowerTokens Video Studio	Open-source (MIT) Video Studio 1.12 (Windows desktop + web) for batch rendering: turns an Excel/CSV shot list into clips with Wan · Seedance · Kling · and more, adds a shared character sheet to every prompt, and resumes unfinished jobs by task ID. Requires a PowerTokens API key.	github.com/PowerTokens/video-studio
VideoPrompt	Free on-device AI prompt library & generators: 103 templates across 21 platforms, text/video/image → prompt (WebGPU, zero upload), 14 languages. Template library open-sourced (CC-BY-4.0) with a pip CLI.	videoprompts.tools · github.com/yangchangjiang/videoprompts
ShortSmash	Prompt-to-published short videos: AI script, neural voiceover in 50 languages, AI-verified stock footage, karaoke captions, one-click publish to 10 platforms.	github.com/hydope/shortsmash
MakeMovie	Self-hosted open-source AI video production platform: source text → script, storyboards, assets, I2V clips, voice cloning (local Qwen3-TTS), score, composed master with subtitles, and EDL/FCPXML handoff to Premiere Pro / DaVinci Resolve.	github.com/DreamOfXM/makemovie
OpenMontage Drama Studio	Open-source (AGPL-3.0), self-hostable web studio that turns a one-line idea into a short drama, short-form video or e-commerce product clip through 7 editable stages (script, characters, scenes, props, storyboard & shots, voice, edit); Seedream for reference images and Seedance for shots, bring your own API keys (paid).	github.com/charliewwdev/OpenMontage · openmontage.dtok.io
VidFare	Free tracker of AI video generation prices in dollars per second of output across APIs and credit plans (9 models, 10 platforms, re-checked daily); each price links to its vendor source; credit-plan conversions are marked as estimates. Includes a deals page with end dates and a credit-plan calculator.	vidfare.com · github.com/SheepinBoots/clipcost
Atlas Cloud	Pay-as-you-go API platform that serves text-to-video and image-to-video models from several vendors (including Seedance, Kling, Wan, Veo, Vidu and Hailuo) behind one API key: submit an async generation request, then poll for the result. Also offers image, audio and LLM models.	atlascloud.ai
Phosphene	Generates video from text and images locally on Apple Silicon Macs using MLX.	github.com/mrbizarro/Phosphene
🎤 Barz Studio	Browser lyric-video editor: timed captions and beat-reactive visuals from your own track; free 720 × 1280 WebM export, with optional automatic lyrics.	barzvideo.com

---

How to Choose

Your Goal	Recommended Tool	Why
Professional filmmaking / VFX	Runway Gen-4.5	Best creative control, reference consistency, Aleph editing
Best motion realism on a budget	Kling 3.0	Top-tier motion, generous free tier, native audio
4K broadcast / cinematic scenes	Veo 3.1	Native 4K, scene extension, lip-sync audio
Multi-shot storytelling	Seedance 2.0	Unified audio-video generation, timeline prompting
Anime / stylized content	PixVerse V6	Character consistency, camera controls, native audio
Corporate training / LMS	Synthesia / Colossyan	SCORM, avatars, quizzes, multilingual
Avatar marketing videos	HeyGen / D-ID	Expressive avatars, fast generation, API access
Real-time creative exploration	Krea AI	64+ models, sub-50ms feedback
Local / open-source deployment	Wan 2.2 / HunyuanVideo 1.5	Apache 2.0; Wan 2.2 offers a 5B model for 24 GB GPUs

---

Open-Source Models & Toolkits

Model	Preview	Org	Size	Code	Min VRAM	Best For
Wan 2.2		Alibaba	27B (14B active) / 5B		24 GB (5B)	Open T2V/I2V, MoE models, speech-to-video support
HunyuanVideo 1.5		Tencent	8.3B		~14 GB	Fast iteration, efficient quality, bilingual
LTX-Video / LTX-2.3		Lightricks	22B		~8 GB	Real-time / 4K generation with native audio
CogVideoX		THUDM / Zhipu AI	2B / 5B		~16 GB	ComfyUI ecosystem, LoRA support, beginners
Open-Sora 2.0		HPC-AI Tech	11B		~24 GB	Full training pipeline, commercial-level quality
Open-Sora-Plan 1.3		PKU-YuanGroup	—		Varies	Open reproduction of Sora capabilities
Mochi 1		Genmo	10B		~24 GB	Realistic physics and motion
Step-Video-T2V		StepFun	—		High	Large open-source T2V model series
Stable Video Diffusion		Stability AI	—		~16 GB	Image-to-video / video foundation model
AnimateDiff		Tsinghua / CUHK	—		~12 GB	Animate personalized T2I models without tuning
Latte		Shanghai AI Lab / NTU	—		~16 GB	Early DiT-based latent diffusion for video
SkyReels V1 / V2		SkyworkAI	1.3B / 14B		~16 GB	Human-centric V1; V2 uses diffusion-forcing for infinite-length film
MAGI-1		Sand AI	4.5B / 24B		Varies	Autoregressive chunk-based generation, up to 4M tokens context
Waver 1.0		ByteDance	12B		~24 GB	Unified T2V/I2V/T2I, 1080p, Cascade Refiner
VideoCrafter2		Tencent AI Lab / CUHK	—		~16 GB	High-quality T2V/I2V diffusion models
ModelScope T2V		Alibaba / DAMO	—		~16 GB	Early open diffusion T2V model (UNet3D)
Allegro / Allegro-TI2V		Rhymes AI	~3B		~9 GB (offload)	Apache 2.0 commercial use, Diffusers integration
DynamiCrafter		CUHK / Tencent AI Lab	—		~16 GB	(Text-)Image-to-video animation from still images
I2VGen-XL		Alibaba / DAMO	—		~16 GB	High-fidelity image-to-video with semantic understanding
VideoComposer		Alibaba / DAMO	—		~16 GB	Compositional video synthesis with motion controllability
Lumiere (unofficial)		lucidrains	—		N/A	PyTorch implementation of Google DeepMind's Lumiere

Quick Start

Get started with the most popular open-source models:

Wan 2.2

git clone https://github.com/Wan-Video/Wan2.2
cd Wan2.2
pip install -r requirements.txt

HunyuanVideo 1.5

git clone https://github.com/Tencent-Hunyuan/HunyuanVideo
cd HunyuanVideo
# Follow the official guide for sampling and I2V inference

CogVideoX

git clone https://github.com/THUDM/CogVideo
cd CogVideo
pip install -r requirements.txt

For detailed inference configs, LoRA fine-tuning, and ComfyUI workflows, see each repository's README.

---

Research Papers

2026

• Efficient Video Diffusion Models: Advancements and Challenges, Shitong Shao. [Paper]
  • First comprehensive survey focused on efficient video diffusion: step distillation, efficient attention, model compression, caching/trajectory optimization.
• Partition the Support, Reconstruct the Residual: Training-Free Sparse Attention for Video Generation and World Models, Pardis Taghavi, Reza Langari, and Gaurav Pandey. [Paper] [Project]
  • Training-free sparse attention using response-coupled partitioning and probe-fitted residual reconstruction, evaluated across four video generation and world models.
• Controllable Video Generation: A Survey [Paper]
  • Systematic review of pose-guided, structure-controlled, and other conditional video generation methods.
• Survey of Video Diffusion Models: Foundations, Implementations, and Applications, Yimu Wang et al. [Paper]
  • Comprehensive taxonomy of diffusion-based video generation, evaluation metrics, industry solutions, and training engineering.
• UNIC: Unified In-Context Video Editing [Paper]
• BRITE: A Benchmark for Reliable and Interpretable T2V Evaluation on Implausible Scenarios [Paper]
• Runway Gen-4.5 Technical Report — Runway (2026) [Runway Help]
• ByteDance Seedance 2.0 — Global launch April 2026 [Overview]
• LTX-2.3: Native 4K Video + Audio Generation — Lightricks (March 2026) [Project]

2025

• Bridging Text and Video Generation: A Survey, Nilay Kumar et al. [Paper]
  • Comprehensive survey on T2V evolution, DiT architectures, datasets, training recipes, and benchmarks.
• Wan 2.2: Mixture-of-Experts Video Generation — Alibaba (July 2025) [Project]
• Wan: Open and Advanced Large-Scale Video Generative Models, Alibaba. [Paper] [Code]
• Training a Commercial-Level Video Generation Model in $200k (Open-Sora 2.0) [Paper] [Code]
• Step-Video-T2V Technical Report: The Practice, Challenges, and Future of Video Foundation Model, Guoqing Ma et al. [Paper] [Code]
• HunyuanVideo: A Systematic Framework For Large Video Generative Models, Weijie Kong et al. [Paper] [Code]
• Fast Video Generation with Sliding Tile Attention, Peiyuan Zhang et al. [Paper]
• Training-Free Efficient Video Generation via Dynamic Token Carving [Paper]
• FastCar: Cache Attentive Replay for Fast Auto-Regressive Video Generation on the Edge [Paper]
• On-Device Sora: Enabling Training-Free Diffusion-Based Text-to-Video Generation for Mobile Devices [Paper]
• SkyReels-A2: Compose Anything in Video Diffusion Transformers [Paper]
• MotionCanvas: Cinematic Shot Design with Controllable Image-to-Video Generation [Paper]
• EasyControl: Adding Efficient and Flexible Control for Diffusion Transformer [Paper]
• MUG-V 10B: High-efficiency Training Pipeline for Large Video Generation Models [Paper]
• VBench-2.0: Advancing Video Generation Benchmark Suite for Intrinsic Faithfulness [Paper] [Code]
• Kandinsky 5.0: A Family of Foundation Models for Image and Video Generation [Paper]
• ARLON: Boosting Diffusion Transformers with Autoregressive Models for Long Video Generation [Paper]
• T2VTextBench: A Human Evaluation Benchmark for Textual Control in Video Generation Models [Paper]
• A Survey of AI-Generated Video Evaluation [Paper]

2024

• Movie Gen: A Cast of Media Foundation Models, Meta. [Paper]
• Lumiere: A Space-Time Diffusion Model for Video Generation, Google DeepMind. [Paper] [Project] [Code]
• Vidu: A Highly Consistent, Dynamic and Skilled Text-to-Video Generator with Diffusion Models, Shengshu / Tsinghua. [Paper] [Code]
• Allegro: Open the Black Box of Commercial-Level Video Generation Model, Rhymes AI. [Paper] [Code]
• DynamiCrafter: Animating Open-domain Images with Video Diffusion Priors, CUHK / Tencent AI Lab. [Paper] [Code]
• Sora: A Review on Background, Technology, Limitations, and Opportunities of Large Vision Models, Yixin Liu et al. [Paper]
• VideoCrafter2: Overcoming Data Limitations for High-Quality Video Diffusion Models, Haoxin Chen et al. [Paper] [Code]
• CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer, Zhuoyi Yang et al. [Paper] [Code]
• Latte: Latent Diffusion Transformer for Video Generation, Xin Ma et al. [Paper] [Code]
• VideoTetris: Towards Compositional Text-to-Video Generation [Paper]
• MagicTime: Time-lapse Video Generation Models as Metamorphic Simulators [Paper]
• T2V-CompBench: A Comprehensive Benchmark for Compositional Text-to-Video Generation [Paper]
• CPA: Camera-pose-awareness Diffusion Transformer for Video Generation [Paper]
• Open-Sora: Democratizing Efficient Video Production for All [Project]
• Open-Sora-Plan: Open-Source Reproduction of Sora [Project]

2023

• Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasets, Andreas Blattmann et al. [Paper]
• ModelScope Text-to-Video Technical Report, Jiayu Wang et al. [Paper]
• I2VGen-XL: High-Quality Image-to-Video Synthesis via Cascaded Diffusion Models, Alibaba / DAMO. [Paper] [Code]
• AnimateDiff: Animate Your Personalized Text-to-Image Diffusion Models without Specific Tuning, Yuwei Guo et al. [Paper] [Code]
• Scalable Diffusion Models with Transformers (DiT), William Peebles & Saining Xie. [Paper] — Foundation for diffusion-transformer architectures used in Sora and successors.
• Video Generation Models as World Simulators, OpenAI (Sora technical report). [Report]
• Text-To-4D Dynamic Scene Generation, Uriel Singer et al. [Paper] [Project]

2022

• Tune-A-Video: One-Shot Tuning of Image Diffusion Models for Text-to-Video Generation, Jay Zhangjie Wu et al. [Paper] [Project] [Code]
• MagicVideo: Efficient Video Generation With Latent Diffusion Models, Daquan Zhou et al. [Paper] [Project]
• Phenaki: Variable Length Video Generation From Open Domain Textual Description, Ruben Villegas et al. [Paper]
• Imagen Video: High Definition Video Generation with Diffusion Models, Jonathan Ho et al. [Paper] [Project]
• Text-driven Video Prediction, Xue Song et al. [Paper]
• Make-A-Video: Text-to-Video Generation without Text-Video Data, Uriel Singer et al. [Paper] [Project] [Code]
• StoryDALL-E: Adapting Pretrained Text-to-Image Transformers for Story Continuation, Adyasha Maharana et al. [Paper] [Code]
• Word-Level Fine-Grained Story Visualization, Bowen Li et al. [Paper] [Code]
• CogVideo: Large-scale Pretraining for Text-to-Video Generation via Transformers, Wenyi Hong et al. [Paper] [Code]
• Show Me What and Tell Me How: Video Synthesis via Multimodal Conditioning, Yogesh Balaji et al. [Paper] [Code] [Project]
• Video Diffusion Models, Jonathan Ho et al. [Paper] [Project]

Before 2021

• Transcript to Video: Efficient Clip Sequencing from Texts, Ligong Han et al. [Paper] [Project]
• GODIVA: Generating Open-DomaIn Videos from nAtural Descriptions, Chenfei Wu et al. [Paper]
• Text2Video: Text-driven Talking-head Video Synthesis with Phonetic Dictionary, Sibo Zhang et al. [Paper]
• TiVGAN: Text to Image to Video Generation With Step-by-Step Evolutionary Generator, Doyeon Kim et al. [Paper]
• Conditional GAN with Discriminative Filter Generation for Text-to-Video Synthesis, Yogesh Balaji et al. [Paper] [Code]
• IRC-GAN: Introspective Recurrent Convolutional GAN for Text-to-video Generation, Kangle Deng et al. [Paper]
• StoryGAN: A Sequential Conditional GAN for Story Visualization, Yitong Li et al. [Paper] [Code]
• Video Generation From Text, Yitong Li et al. [Paper]
• To create what you tell: Generating videos from captions, Yingwei Pan et al. [Paper]

---

Datasets & Benchmarks

Datasets

Dataset	Scale	Highlights	Link
WebVid-10M	10.7M clips	Large-scale text-video pairs scraped from the web	m-bain/webvid
InternVid	7M+ clips	High-quality video-text dataset with multimodal annotations	OpenGVLab/InternVid
HD-VILA-100M	100M clips	High-resolution long-form videos with dense captions	microsoft/HQD
Panda-70M	70M clips	Large dataset of high-quality video-caption pairs	snap-research/Panda-70M
Vript	420K clips	Ultra-detailed captions (~145 words), shot type and camera movement	Vript
MiraData	798K clips	Long-duration videos (avg 72s), 1080p, 318-word captions	MiraData
OpenVid-1M	1M clips	High-quality, diverse scenarios, ~98-word captions	OpenVid-1M
CI-VID	1M clips / 717K seqs	Coherent interleaved text-video dataset	CI-VID
HD-VG-130M	130M clips	High-definition, watermark-free video-text pairs	HD-VG-130M
VidProM	—	Large prompt-gallery dataset for video generation	VidProM

Benchmarks

Benchmark	Focus	Link
VBench / VBench-2.0	Comprehensive video generation evaluation suite	Vchitect/VBench
T2V-CompBench	Compositional text-to-video generation	T2V-CompBench
T2VTextBench	Human evaluation of textual control in video generation	T2VTextBench
ChronoMagic-Bench	Metamorphic / time-lapse text-to-video evaluation	ChronoMagic-Bench
BRITE	Reliable T2V evaluation on implausible scenarios	arXiv:2605.00873
VideoEval	Low-cost evaluation of video foundation models	VideoEval
VideoScore2	Think-before-you-score reward model / metric for generated video: visual quality, T2V alignment & physical consistency with chain-of-thought (ships VideoScore-Bench-v2)	Paper · Code

---

Contributing

Contributions are welcome! Please open a pull request to add new products, papers, models, datasets, or benchmarks. Keep entries concise and include both a paper link and a code/project link when available.

If you have any questions, feel free to contact jianzhnie.

---

License

Awesome-Text-To-Video is released under the Apache 2.0 license.

---

References

This list was compiled and updated using information from the following sources.

Product Websites

• OpenAI Sora
• Google DeepMind Veo
• Runway • Runway Help: Creating with Gen-4 Video • Runway Review 2026
• Luma Dream Machine
• Kling AI
• Hailuo AI
• Pika
• Krea AI
• ByteDance Seedance • Seedance 2.0 Global Release Guide • Seedance 2.0 Review
• PixVerse
• Morphic / Morph Studio
• Hedra
• Vidnoz AI
• Steve.AI
• Vidu
• Lumiere Project Page
• Synthesia
• HeyGen
• DeepBrain AI
• Colossyan
• Elai.io
• D-ID
• Hour One
• Creatify
• InVideo AI
• Magic Hour
• videos.social
• Unsora
• Katechon Arc
• Pixwit
• RefVid
• Cicadas.ai

Open-Source Repositories

• Wan 2.1
• HunyuanVideo
• CogVideo
• Open-Sora
• Open-Sora-Plan
• LTX-Video
• Mochi
• Step-Video-T2V
• SkyReels-V2
• MAGI-1
• Waver
• VideoCrafter
• VGen / ModelScope T2V
• Allegro
• DynamiCrafter
• I2VGen-XL
• VideoComposer
• Lumiere PyTorch
• AnimateDiff
• Latte

Datasets, Benchmarks & Surveys

• VBench
• T2V-CompBench
• T2VTextBench
• ChronoMagic-Bench
• VideoEval
• arXiv papers cited in the Research Papers section.