whatmodel.sh · find the best video models you can run

Best video models to run locally

A leaderboard of the best video models you can run on your own computer. MiniMax H3 is #1 right now, followed by Cosmos3-Super and Wan 2.2 I2V A14B. Models are ranked by one score that combines 55 published benchmarks. Each model shows the download that fits and the command to run it.

51 models · scores combine 55 benchmarks · updated how it works · find a leaderboard · benchmark your machine
task[all tasks][text to video][image to video][video editing]
Filter by your hardware and budget · local only · reset
+ more

Detect your exact chip or graphics card:

curl -G whatmodel.sh/video -d max=0 --data-urlencode "mac=$(sysctl -n hw.memsize machdep.cpu.brand_string)"Mac

curl -G whatmodel.sh/video -d max=0 --data-urlencode "ram=$(awk '/MemT/{print $2}' /proc/meminfo)" --data-urlencode "nv=$(nvidia-smi --query-gpu=name,memory.total --format=csv,noheader 2>/dev/null)"Linux

curl.exe -G whatmodel.sh/video -d max=0 --data-urlencode "ram=$((gcim Win32_ComputerSystem).TotalPhysicalMemory)" --data-urlencode "nv=$(nvidia-smi --query-gpu=name,memory.total --format=csv,noheader 2>$null)"Windows

$ terminal

detecting your OS…

Run curl "whatmodel.sh/video?max=0"to get this list in your terminal (add &format=json for agents).

ram[8][16][24][32][48][64][96][128][192][256]GB
gpu[apple silicon][graphics card][none]
price[local][$0.02][$0.05][$0.1][$0.25][any]max per sec

detecting your OS…

Score vs memory to run

#modelscore↓where↓speed↓RAM↓get itsources
#modelscore↓where↓speed↓RAM↓get itsources
6MiniMax H3 MiniMax78±1local~40GBhow to run › LMArena leaderboard datasetRapidata/world-model-physics published dataset-card rankingsRapidata/relative-camera-movement-human-preference-900k published dataset-card rankingsVGI-Bench: Probing Visual Intelligence in Video Generation ModelsRapidata/image-to-video-realism published dataset-card rankingsTraining Object Permanence in World Models (WROP exam; arXiv:2609.28654v1)+2
38Cosmos3-Super NVIDIA58±4*†local~160GBhow to run › Cosmos 3: Omnimodal World Models for Physical AI (arXiv:2606.02800v4)
42Wan 2.2 I2V A14B Alibaba56±2local~31GBhow to run › WorldReasonBench: Human-Aligned Stress Testing of Video Generators as Future World-State Predictors (arXiv v1)CRONOS: Benchmarking Counterfactual Physical Consistency in Video Models (arXiv:2605.23699v1), Table 1PAI-Bench: A Comprehensive Benchmark For Physical AI (arXiv:2512.01989v1)MindWorldBench: Evaluating Mental-State-to-Behavior Reasoning in Image-to-Video Generation (arXiv:2609.39147v1)Cosmos 3: Omnimodal World Models for Physical AI (arXiv:2606.02800v4)VGI-Bench: Probing Visual Intelligence in Video Generation Models+2
53LTX-2.5 Lightricks53±6*local~14GBhow to run › Kandinsky 6.0 Video: Foundation Models for Synchronized Video and Audio Generation (arXiv:2610.05608v1)Datapoint Video Bench leaderboard (datapointai/text-to-video-human-preferences-326k dataset card)
55HunyuanVideo 1.5 (I2V) Tencent53±2local~66GBhow to run › WorldReasonBench: Human-Aligned Stress Testing of Video Generators as Future World-State Predictors (arXiv v1)MindWorldBench: Evaluating Mental-State-to-Behavior Reasoning in Image-to-Video Generation (arXiv:2609.39147v1)Cosmos 3: Omnimodal World Models for Physical AI (arXiv:2606.02800v4)VGI-Bench: Probing Visual Intelligence in Video Generation ModelsVideoArgus: Agentic Rubric-Grounded Unified Evaluation for Video Generation and Editing (arXiv:2608.05485v1)MechVerse: Evaluating Physical Motion Consistency in Video Generation Models (arXiv:2605.14843v1)+2
57Kandinsky 5.0 Pro Sber52±2*local~76GBhow to run › LMArena leaderboard datasetHumanScore: Benchmarking Human Motions in Generated Videos
59HunyuanVideo 1.5 Tencent52±2local~20GBhow to run › LMArena leaderboard datasetCosmos 3: Omnimodal World Models for Physical AI (arXiv:2606.02800v4)WorldJen: An End-to-End Multi-Dimensional Benchmark for Generative Video ModelsVideoArgus: Agentic Rubric-Grounded Unified Evaluation for Video Generation and Editing (arXiv:2608.05485v1)Sci-VBench: Evaluating Knowledge- and Reasoning-Intensive Video Generation in Science DomainsHumanScore: Benchmarking Human Motions in Generated Videos
60Cosmos3-Nano NVIDIA52±3local~42GBhow to run › Cosmos 3: Omnimodal World Models for Physical AI (arXiv:2606.02800v4)Rapidata/relative-camera-movement-human-preference-900k published dataset-card rankingsRapidata/image-to-video-realism published dataset-card rankingsSci-VBench: Evaluating Knowledge- and Reasoning-Intensive Video Generation in Science DomainsPhysics-IQ Verified
63LongCat-Video Meituan51±2local~100GBhow to run › WorldReasonBench: Human-Aligned Stress Testing of Video Generators as Future World-State Predictors (arXiv v1)MindWorldBench: Evaluating Mental-State-to-Behavior Reasoning in Image-to-Video Generation (arXiv:2609.39147v1)VideoArgus: Agentic Rubric-Grounded Unified Evaluation for Video Generation and Editing (arXiv:2608.05485v1)Sci-VBench: Evaluating Knowledge- and Reasoning-Intensive Video Generation in Science Domains
64Kandinsky 6.0 Video Lite Sber51±7*†local~34GBhow to run › Kandinsky 6.0 Video: Foundation Models for Synchronized Video and Audio Generation (arXiv:2610.05608v1)
67Wan 2.2 A14B Alibaba50±2local~31GBhow to run › LMArena leaderboard datasetCosmos 3: Omnimodal World Models for Physical AI (arXiv:2606.02800v4)WorldJen: An End-to-End Multi-Dimensional Benchmark for Generative Video ModelsVideoVerse: Does Your T2V Generator Have World Model Capability to Synthesize Videos?VideoArgus: Agentic Rubric-Grounded Unified Evaluation for Video Generation and Editing (arXiv:2608.05485v1)
70LTX-2 Lightricks49±2local~120GBhow to run › LMArena leaderboard datasetWorldJen: An End-to-End Multi-Dimensional Benchmark for Generative Video ModelsPhyGround: Benchmarking Physical Reasoning in Generative World Models (arXiv:2605.10806v1)
71Open-Sora Plan v1.3 (T2V) Peking University (PKU-YuanGroup)49±7*local~24GBhow to run › WorldModelBench: Judging Video Generation Models As World Models
77Kandinsky 6.0 Video Pro Sber48±7*†local~109GBhow to run › Kandinsky 6.0 Video: Foundation Models for Synchronized Video and Audio Generation (arXiv:2610.05608v1)
80Wan 2.1 14B Alibaba47±2local~20GBhow to run › VideoPhy / VideoPhy-2 official repository leaderboardsSkyReels-V2: Infinite-length Film Generative ModelCosmos 3: Omnimodal World Models for Physical AI (arXiv:2606.02800v4)T2VPhysBench: A First-Principles Benchmark for Physical Consistency in Text-to-Video GenerationWan-AI/Wan2.1-T2V-14B model cardAlibaba Wan technical report (arXiv:2503.20314v2), VBench table (own models)
81Wan 2.2 TI2V 5B Alibaba47±2local~6.5GBhow to run › Cosmos 3: Omnimodal World Models for Physical AI (arXiv:2606.02800v4)VABench: A Comprehensive Benchmark for Audio-Video Generation (arXiv:2512.09299v2)Target-Bench: Can Video World Models Achieve Mapless Path Planning with Semantic Targets?VideoArgus: Agentic Rubric-Grounded Unified Evaluation for Video Generation and Editing (arXiv:2608.05485v1)Sci-VBench: Evaluating Knowledge- and Reasoning-Intensive Video Generation in Science DomainsPAI-Bench: A Comprehensive Benchmark For Physical AI (arXiv:2512.01989v1)+3
82Kandinsky 5.0 T2V Lite Sber47±2*local~29GBhow to run › LMArena leaderboard dataset
83LTX-2 19B Distilled Lightricks47±8*local~41GBhow to run › Sci-VBench: Evaluating Knowledge- and Reasoning-Intensive Video Generation in Science Domains
85CogVideoX-5B-I2V Zhipu AI (THUDM)46±4local~13GBhow to run › MindWorldBench: Evaluating Mental-State-to-Behavior Reasoning in Image-to-Video Generation (arXiv:2609.39147v1)VideoArgus: Agentic Rubric-Grounded Unified Evaluation for Video Generation and Editing (arXiv:2608.05485v1)WorldModelBench: Judging Video Generation Models As World ModelsMechVerse: Evaluating Physical Motion Consistency in Video Generation Models (arXiv:2605.14843v1)
86Cosmos-1.0-Diffusion-7B-Text2World NVIDIA46±3local~30GBhow to run › VideoPhy / VideoPhy-2 official repository leaderboardsCosmos World Foundation Model Platform for Physical AI (arXiv:2501.03575v3)
90HunyuanVideo-I2V Tencent45±4local~57GBhow to run › PAI-Bench: A Comprehensive Benchmark For Physical AI (arXiv:2512.01989v1)SkyReels-V2: Infinite-length Film Generative Model
91CogVideoX-5B Zhipu AI (THUDM)44±2local~13GBhow to run › VideoPhy / VideoPhy-2 official repository leaderboardsGenAI-Arena (TIGER-Lab) text-to-video leaderboardVideoArgus: Agentic Rubric-Grounded Unified Evaluation for Video Generation and Editing (arXiv:2608.05485v1)WorldModelBench: Judging Video Generation Models As World ModelsHumanScore: Benchmarking Human Motions in Generated Videos
95Open-Sora 2.0 HPC-AI Tech42±5*local~55GBhow to run › VideoVerse: Does Your T2V Generator Have World Model Capability to Synthesize Videos?SkyReels-V2: Infinite-length Film Generative Model
96Pyramid Flow (SD3) Peking University / Kuaishou42±3*local~8GBhow to run › GenAI-Arena (TIGER-Lab) text-to-video leaderboard
99Wan 2.1 T2V 1.3B Alibaba40±4local~8.6GBhow to run › WorldJen: An End-to-End Multi-Dimensional Benchmark for Generative Video ModelsAlibaba Wan technical report (arXiv:2503.20314v2), VBench table (own models)Training-Free Efficient Video Generation via Dynamic Token Carving
100Mochi 1 Genmo40±2local~6.5GBhow to run › LMArena leaderboard datasetGenAI-Arena (TIGER-Lab) text-to-video leaderboardVideoPhy / VideoPhy-2 official repository leaderboardsWan-AI/Wan2.1-T2V-14B model cardT2VPhysBench: A First-Principles Benchmark for Physical Consistency in Text-to-Video GenerationWorldModelBench: Judging Video Generation Models As World Models
102Open-Sora Plan v1.3 (I2V) Peking University (PKU-YuanGroup)39±7*local~24GBhow to run › WorldModelBench: Judging Video Generation Models As World Models
105HunyuanVideo Tencent39±3local~28GBhow to run › VideoPhy / VideoPhy-2 official repository leaderboardsSkyReels-V2: Infinite-length Film Generative ModelVideoVerse: Does Your T2V Generator Have World Model Capability to Synthesize Videos?Training-Free Efficient Video Generation via Dynamic Token Carving
107T2V-Turbo-v2 UC Santa Barbara38±5*local~9.4GBhow to run › T2V-Turbo-v2: Enhancing Video Generation Model Post-Training through Data, Reward, and Conditional Guidance Design (arXiv 2410.05677v3)WorldModelBench: Judging Video Generation Models As World Models
109T2V-Turbo (VC2) UC Santa Barbara38±3†local~9.4GBhow to run › GenAI-Arena (TIGER-Lab) text-to-video leaderboardT2V-Turbo: Breaking the Quality Bottleneck of Video Consistency Model with Mixed Reward Feedback (arXiv:2405.18750v2)T2V-Turbo-v2: Enhancing Video Generation Model Post-Training through Data, Reward, and Conditional Guidance Design (arXiv 2410.05677v3)
112LTX-2.3 Lightricks36±3local~34GBhow to run › WorldReasonBench: Human-Aligned Stress Testing of Video Generators as Future World-State Predictors (arXiv v1)MindWorldBench: Evaluating Mental-State-to-Behavior Reasoning in Image-to-Video Generation (arXiv:2609.39147v1)Sci-VBench: Evaluating Knowledge- and Reasoning-Intensive Video Generation in Science DomainsPhyGround: Benchmarking Physical Reasoning in Generative World Models (arXiv:2605.10806v1)
113CogVideoX1.5-5B Zhipu AI (THUDM)36±2local~38GBhow to run › LOVE: Benchmarking and Evaluating Text-to-Video Generation and Video-to-Text Interpretation (AIGVE-60K)VideoVerse: Does Your T2V Generator Have World Model Capability to Synthesize Videos?VideoArgus: Agentic Rubric-Grounded Unified Evaluation for Video Generation and Editing (arXiv:2608.05485v1)SkyReels-V2: Infinite-length Film Generative ModelSci-VBench: Evaluating Knowledge- and Reasoning-Intensive Video Generation in Science Domains
114CogVideoX-2B Zhipu AI (THUDM)36±2local~5.3GBhow to run › VideoPhy / VideoPhy-2 official repository leaderboardsGenAI-Arena (TIGER-Lab) text-to-video leaderboardChronoMagic-Bench: A Benchmark for Metamorphic Evaluation of Text-to-Time-lapse Video Generation
116Allegro Rhymes AI35±2local~9.3GBhow to run › GenAI-Arena (TIGER-Lab) text-to-video leaderboardLOVE: Benchmarking and Evaluating Text-to-Video Generation and Video-to-Text Interpretation (AIGVE-60K)
118VideoCrafter2 Tencent AI Lab34±2local~9.4GBhow to run › VideoPhy / VideoPhy-2 official repository leaderboardsGenAI-Arena (TIGER-Lab) text-to-video leaderboardLOVE: Benchmarking and Evaluating Text-to-Video Generation and Video-to-Text Interpretation (AIGVE-60K)ChronoMagic-Bench: A Benchmark for Metamorphic Evaluation of Text-to-Time-lapse Video Generation
121Open-Sora 1.1 HPC-AI Tech32±7*local~14GBhow to run › ChronoMagic-Bench: A Benchmark for Metamorphic Evaluation of Text-to-Time-lapse Video Generation
122LaVie (base T2V) Shanghai AI Lab (Vchitect)32±2local~4.9GBhow to run › GenAI-Arena (TIGER-Lab) text-to-video leaderboardLOVE: Benchmarking and Evaluating Text-to-Video Generation and Video-to-Text Interpretation (AIGVE-60K)VideoPhy / VideoPhy-2 official repository leaderboardsChronoMagic-Bench: A Benchmark for Metamorphic Evaluation of Text-to-Time-lapse Video Generation
123LTX-Video 0.9 (2B) Lightricks32±3*local~5.3GBhow to run › GenAI-Arena (TIGER-Lab) text-to-video leaderboard
125NOVA-0.6B T2V BAAI29±3*local~9.3GBhow to run › LOVE: Benchmarking and Evaluating Text-to-Video Generation and Video-to-Text Interpretation (AIGVE-60K)
126ZeroScope v2 576w cerspense27±3local~7.9GBhow to run › VideoPhy / VideoPhy-2 official repository leaderboardsChronoMagic-Bench: A Benchmark for Metamorphic Evaluation of Text-to-Time-lapse Video GenerationFETV: A Benchmark for Fine-Grained Evaluation of Open-Domain Text-to-Video Generation (arXiv:2311.01813v3)
127Hotshot-XL Hotshot27±2local~10GBhow to run › EvalCrafter: original 2023 text-to-video evaluationsLOVE: Benchmarking and Evaluating Text-to-Video Generation and Video-to-Text Interpretation (AIGVE-60K)
130Open-Sora 1.2 HPC-AI Tech27±3local~39GBhow to run › GenAI-Arena (TIGER-Lab) text-to-video leaderboardWorldModelBench: Judging Video Generation Models As World ModelsChronoMagic-Bench: A Benchmark for Metamorphic Evaluation of Text-to-Time-lapse Video Generation
133VideoCrafter1 T2V Tencent AI Lab / HKUST26±2local~8.1GBhow to run › EvalCrafter: original 2023 text-to-video evaluations
134Show-1 Show Lab / NUS25±2local~83GBhow to run › EvalCrafter: original 2023 text-to-video evaluations
135LVDM (text-to-video) HKUST / Tencent AI Lab25±3*local~6.1GBhow to run › LOVE: Benchmarking and Evaluating Text-to-Video Generation and Video-to-Text Interpretation (AIGVE-60K)
136ModelScope T2V 1.7B Alibaba DAMO Academy25±2local~4.6GBhow to run › EvalCrafter: original 2023 text-to-video evaluationsGenAI-Arena (TIGER-Lab) text-to-video leaderboardLOVE: Benchmarking and Evaluating Text-to-Video Generation and Video-to-Text Interpretation (AIGVE-60K)FETV: A Benchmark for Fine-Grained Evaluation of Open-Domain Text-to-Video Generation (arXiv:2311.01813v3)ChronoMagic-Bench: A Benchmark for Metamorphic Evaluation of Text-to-Time-lapse Video Generation
138ZeroScope v2 XL (576w → XL pipeline) cerspense24±2local~15GBhow to run › EvalCrafter: original 2023 text-to-video evaluations
139CogVideo 9B (original) Tsinghua / Zhipu23±7*local~52GBhow to run › FETV: A Benchmark for Fine-Grained Evaluation of Open-Domain Text-to-Video Generation (arXiv:2311.01813v3)
—AnimateDiff SD 1.5 v2 motion adapter AnimateDiff authorsno reviewed resultslocal~7.8GBhow to run ›
—Stable Video Diffusion (SVD, 14 frames) Stability AIno reviewed resultslocal~8GBhow to run ›
—Stable Video Diffusion XT (25 frames) Stability AIno reviewed resultslocal~8GBhow to run ›

+ 91 more don't fit your machine or budget · show them

142 of 142 catalogue models listed here · sources and licences

* new or thinly tested: ranked on too few benchmark results · † only the maker’s own numbers so far · ~RAM is estimated; full context may need more memory