Ranking de texto a video (Con audio)Artificial Analysis
Minimax H3 Max (post-trained by fal), Agnes-Video-2.5, LTX-2.5 Fast, LTX-2.5 Pro, Wan 3.0
Range | Creator | Model | Elo | 95% CI | Samples | Lanzamiento | Precio de la API 1 | |
|---|---|---|---|---|---|---|---|---|
| 1 | 1-2 |  Alibaba | Wan 3.0 | 1240 | -9/9 | 5752 | ago 2026 | 12,00 US$ /min |
| 2 | 1-2 |  Google | Gemini Omni Flash | 1237 | -6/6 | 16.380 | may 2026 | 6,00 US$ /min |
| 3 | 2-4 |  Fal | Minimax H3 Max (post-trained by fal) | 1230 | -9/9 | 5492 | ago 2026 | 2,40 US$ /min |
| 4 | 3-5 |  MiniMax | MiniMax H3  Open Weights | 1225 | -7/7 | 8864 | jul 2026 | 7,80 US$ /min |
| 5 | 4-5 |  ByteDance Seed | Dreamina Seedance 2.0 720p | 1220 | -6/6 | 22.625 | mar 2026 | 9,07 US$ /min |
| 6 | 6 | Alibaba | Wan2.7-260612 | 1158 | -6/6 | 18.336 | jun 2026 | 9,00 US$ /min |
| 7 | 7 | Alibaba-ATH | HappyHorse-1.1 | 1144 | -6/6 | 18.463 | jun 2026 | 9,90 US$ /min |
| 8 | 8 | Alibaba-ATH | HappyHorse-1.0 | 1124 | -6/6 | 9388 | abr 2026 | 13,20 US$ /min |
| 9 | 9-10 |  Sand.ai | MAGI-2 Preview Open Weights | 1114 | -9/9 | 5611 | ago 2026 | Próximamente |
| 10 | 10-12 |  KlingAI | Kling 3.0 1080p (Pro) | 1108 | -5/5 | 20.843 | feb 2026 | 20,16 US$ /min |
| 11 | 10-14 | Alibaba | Wan 2.7 | 1104 | -7/7 | 5991 | abr 2026 | 9,00 US$ /min |
| 12 | 10-14 |  Skywork AI | SkyReels V4 | 1103 | -7/7 | 6511 | mar 2026 | 21,00 US$ /min |
| 13 | 11-14 | KlingAI | Kling 3.0 720p (Standard) | 1101 | -5/5 | 20.031 | feb 2026 | 15,12 US$ /min |
| 14 | 11-17 |  OpenAI | Sora 2 (December) | 1098 | -9/9 | 4838 | dic 2025 | 6,00 US$ /min |
| 15 | 15-20 | KlingAI | Kling 3.0 Omni 1080p (Pro) | 1090 | -6/6 | 8873 | feb 2026 | 16,80 US$ /min |
| 16 | 15-20 | KlingAI | Kling 3.0 Omni 720p (Standard) | 1090 | -6/6 | 8635 | feb 2026 | 13,44 US$ /min |
| 17 | 15-20 | Google | Veo 3.1 | 1090 | -6/6 | 8978 | ene 2026 | 24,00 US$ /min |
| 18 | 15-21 | Google | Veo 3.1 Lite | 1087 | -6/6 | 8421 | mar 2026 | 4,80 US$ /min |
| 19 | 15-21 |  Sapiens AI | Agnes-Video-2.5 | 1085 | -8/8 | 6031 | ago 2026 | 1,50 US$ /min |
| 20 | 18-21 | Google | Veo 3.1 Fast | 1085 | -5/5 | 18.966 | ene 2026 | 9,00 US$ /min |
| 21 | 15-23 | OpenAI | Sora 2 Pro | 1081 | -9/9 | 4871 | sept 2025 | 30,00 US$ /min |
| 22 | 21-23 |  PixVerse | PixVerse V6 | 1077 | -6/6 | 9954 | mar 2026 | 6,90 US$ /min |
| 23 | 22-23 |  Vidu | Vidu Q3 Pro | 1075 | -5/5 | 18.863 | ene 2026 | 9,60 US$ /min |
| 24 | 22-25 |  Lightricks | LTX-2.5 Fast Open Weights | 1069 | -9/9 | 2991 | ago 2026 | 7,80 US$ /min |
| 25 | 25-26 |  SpaceXAI | grok-imagine-video | 1062 | -6/6 | 17.858 | ene 2026 | 4,20 US$ /min |
| 26 | 25-26 | Lightricks | LTX-2.5 Pro Open Weights | 1058 | -9/9 | 2972 | ago 2026 | 10,20 US$ /min |
| 27 | 27-28 | Alibaba | Wan 2.6 | 1026 | -7/7 | 7379 | dic 2025 | 9,00 US$ /min |
| 28 | 27-28 | Vidu | Vidu Q3 Turbo | 1022 | -9/9 | 3239 | feb 2026 | 3,90 US$ /min |
| 29 | 29 | ByteDance Seed | Seedance 1.5 pro | 1000 | 0/0 | 12.416 | dic 2025 | 11,86 US$ /min |
| 30 | 30-31 | KlingAI | Kling 2.6 Pro (January) | 983 | -7/7 | 7914 | ene 2026 | 8,40 US$ /min |
| 31 | 30-31 | Lightricks | LTX-2.3 Fast Open Weights | 979 | -6/6 | 11.872 | mar 2026 | 2,40 US$ /min |
| 32 | 32 | Lightricks | LTX-2.3 Pro Open Weights | 961 | -6/6 | 11.250 | mar 2026 | 4,80 US$ /min |
| 33 | 33 | PixVerse | PixVerse V5.6 | 950 | -7/7 | 6359 | feb 2026 | Próximamente |
| 34 | 34 | Lightricks | LTX-2 Fast Open Weights | 941 | -8/8 | 6038 | oct 2025 | 2,40 US$ /min |
| 35 | 35-36 | Lightricks | LTX-2 Pro Open Weights | 919 | -8/8 | 6017 | oct 2025 | 3,60 US$ /min |
| 36 | 35-36 | Sapiens AI | Agnes-Video-V2.0 | 918 | -8/8 | 5307 | may 2026 | 0,30 US$ /min |
1 El precio de la API refleja el costo de generar 1 minuto de video a 1080p en la API del creador del modelo con la configuración predeterminada
Preguntas frecuentes
Los mejores modelos de texto a video con audio por puntuación Elo son: 1. Wan 3.0 (Elo 1240), 2. Gemini Omni Flash (Elo 1237), 3. Minimax H3 Max (post-trained by fal) (Elo 1230), 4. MiniMax H3 (Elo 1225) y 5. Dreamina Seedance 2.0 720p (Elo 1220). La clasificación se basa en votaciones de usuarios a ciegas en la Arena de video de Artificial Analysis.
Wan 3.0 lidera actualmente la Arena de texto a video de Artificial Analysis (sin audio), con una puntuación Elo de 1335.
MiniMax H3 lidera actualmente los modelos de texto a video sin audio y con pesos abiertos en la Arena de texto a video de Artificial Analysis, con una puntuación Elo de 1302, seguido de LTX-2.5 Fast (Elo 1216) y LTX-2.5 Pro (Elo 1206).