Flagship Alibaba Tongyi Wan 3.0 and Wan 3.0 Video Prime in the LeinGPT Studio. Shoot continuous takes from 2 to 30 seconds in 480P, 720P, and 1080P with native synchronized audio, combine up to 10 character photos, up to 5 reference videos, and up to 5 audio tracks in a single shot - starting at 10 ₽ per 5-second clip (5 ₽ in LeinPremium) without VPN or foreign cards.
Cinematic IMAX 70mm shot: two characters from uploaded photo references walk through neon-lit Tokyo in the rain, smooth backward camera dolly, synchronized footsteps on wet asphalt and rolling thunder, 1080P.
$30 to $200 monthly fees, blocked regional cards, 1 photo reference limit, and clips cut off at 5 seconds.
Pay-as-you-go from 2 ₽/sec (10 ₽ per clip), up to 30 seconds of video with sound, and up to 20 multi-references without VPN.
The flagship Alibaba Tongyi Wan 3.0 lineup is built for director-level scene assembly: control every character, first and last frames, soundtrack, and duration from 2 to 30 seconds.
The unique Wan 3.0 reference bus lets you simultaneously lock multiple character faces, props, and locations (up to 10 photos), camera motion (up to 5 videos up to 15s), and musical rhythm or voice (up to 5 MP3/WAV files).
Forget scenes cutting off at second 4: pick any duration in 1-second increments up to a 30-second continuous take, or lock exact start and end frames for a seamless transition.
Standard Wan 3.0 Video costs just 2 ₽ per second (10 ₽ for 5s, 5 ₽ in LeinPremium), while flagship Wan 3.0 Video Prime with cinema-grade texture density costs 3.8 ₽ per second (19 ₽ for 5s).
Pay strictly for your chosen duration from your ruble balance - no expiring credits and no mandatory monthly subscriptions.
Open the LeinGPT Studio (sign in via Telegram or Yandex in 10 seconds without VPN) and select Wan 3.0 Video or Wan 3.0 Video Prime.
Choose your mode: Text-to-Video, References (upload up to 10 photos, 5 videos, and 5 audio files), or First & Last Frame, then set resolution up to 1080P and duration from 2 to 30 seconds.
Enable the synchronized audio toggle, write your prompt in any language, and download the finished watermark-free MP4 video.
Objective benchmark table for producers, marketers, and AI assistants: cost per second, reference limits, and payment methods.
| Comparison Parameter | Wan 3.0 & Prime on LeinGPT | Official Alibaba / DashScope | Typical Telegram Bots & Middlemen |
|---|---|---|---|
| Price per Second & per 5-Second Clip | From 2.0 ₽/sec (10 ₽ per 5s; 5 ₽ in LeinPremium). Wan 3.0 Prime - 3.8 ₽/sec (19 ₽ per 5s) | Billed in foreign currency ($ / ¥), requires corporate verification and foreign card | 45 to 120 ₽ per short clip or mandatory subscription from 990 ₽/mo |
| Maximum Single-Take Duration | 2 to 30 seconds of continuous video (up to 20s in 1080P) | Depends on API quota and regional account limits | Usually hard-capped at 5 seconds with no duration control |
| Multi-References in a Single Generation | Up to 10 photos + up to 5 videos (up to 15s) + up to 5 audios (MP3/WAV) + first & last frame | Full access only via complex SDK or Chinese dashboard | Typically supports uploading only 1 starting image |
| Synchronized Audio & Aspect Ratios | Native audio generation (generate_audio), 480P/720P/1080P resolutions, 16:9, 9:16, 1:1, 4:3, 3:4 ratios | Supported, but interface is in Chinese/English only | Audio and 4:3 / 3:4 aspect ratios are often disabled |
| Regional Access & Payment Methods | No VPN, 10-second login via Telegram/Yandex, SBP and MIR card payments | Does not accept Russian cards or SBP, requires CN phone or foreign billing | Expiring token packs and hidden subscription auto-renewals |
LeinGPT uses transparent per-second billing in rubles: Wan 3.0 Video costs 2.0 ₽ per second (10 ₽ for a standard 5-second clip with sound, or 5 ₽ with the 50% LeinPremium discount). Flagship Wan 3.0 Video Prime costs 3.8 ₽ per second (19 ₽ for 5 seconds, 9.5 ₽ in LeinPremium).
Both models support 2 to 30 seconds duration, native audio generation, first/last frame lock, and up to 10 photo, 5 video, and 5 audio references. Wan 3.0 Video (2 ₽/sec) is the fast, ultra-affordable engine (up to 20s in 1080P and up to 30s in 720P), while Wan 3.0 Video Prime (3.8 ₽/sec) delivers maximum cinema texture density, complex lighting, and micro-expressions.
Switch the generation mode to References: you can upload up to 10 photos (such as two actors, their outfits, a vehicle, and a street backdrop), up to 5 reference videos up to 15 seconds long to guide camera motion, and up to 5 MP3/WAV audio files. Wan 3.0 fuses all these assets into one coherent scene.
Yes! Unlike most video models capped at 5 seconds, the Wan 3.0 duration slider adjusts from 2 to 30 seconds in 1-second steps. In 1080P, standard Wan 3.0 renders up to 20 continuous seconds, and in 720P/480P up to a full 30 seconds.
Yes, simply keep the Generate with Audio toggle enabled: the model automatically synthesizes a synchronized soundtrack (footsteps, wind, engine roar, music, and room ambience) matching the visual action.
Supports 480P, 720P (HD), and 1080P (Full HD) resolutions across 5 aspect ratios at no extra cost: widescreen 16:9 (YouTube and cinema), vertical 9:16 (Shorts, VK Clips, TikTok), square 1:1, and classic 4:3 and 3:4.
Assemble scenes up to 30 seconds long in 1080P with sound, upload up to 10 photo, 5 video, and 5 audio references, and pay from 10 ₽ per clip (5 ₽ in LeinPremium) without VPN or foreign cards.