Seedance 2.5 R2V Basic
seedance-2-5-reference-to-video-basic- 🏢 ByteDance's flagship video family, reference-to-video ("R2V") route, released August 2026.
- 🆕 Joint audio-video generation aimed at 30-second storytelling in a single pass.
- 📏 Clips can be extended twice for longer, more complete narratives.
- 🔧 R2V drives camera movement, pacing, blocking and shot scale from reference clips.
- 👁️ Improved lighting control derives physically plausible light direction, temperature and shadows.
- 🎯 Supports local video editing alongside generation.
- 🌐 ByteDance frames it as moving from clip-level outputs to full creative workflows.
ByteDance is a Chinese internet technology company headquartered in Beijing's Haidian district, with an associated variable-interest entity incorporated in the Cayman Islands. Best known as a consumer internet operator at enormous scale, it has also become a significant…
Explore 18 more models by ByteDance →Seedance 2.5 R2V is the reference-to-video endpoint of ByteDance's Seedance 2.5 generation, sitting alongside the text-to-video Seedance 2.5 and image-to-video Seedance 2.5 routes. ByteDance describes the model as a next-generation audio-video joint generation system built for 30-second storytelling with precise reference control and editing capabilities, where a single generation can be extended twice for longer sequences.
R2V is the route that exposes the most explicit reference-based control of the three. Rather than a single prompt or hero image, it takes labelled reference assets and lets the prompt assign roles to each: ByteDance's own examples reference camera movement, pacing, shot-size transitions, subject trajectory and blocking from one clip while pulling materials, lighting, colour and atmosphere from a separate image.
Compared with Seedance 2.0 R2V and the lighter Seedance 2.0 Mini R2V, the emphasis shifts from short multi-shot clips to longer native takes with fewer stitched segments. ByteDance highlights smoother, more consistent motion and more realistic visuals, plus a specific upgrade in lighting control: spatial information from a reference (for example a clay render) is used to generate light that follows physical laws in direction, colour temperature, intensity and shadow projection. ByteDance frames 2.5 as a step from clip-level outputs toward complete creative workflows.
Practically, that means locking a character, set and palette across a longer take, then making targeted local edits. For still frames, the Seedream V5 Pro image models cover the same family's image side.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
Daily usage · Seedance 2.5 R2V Basic · Organic sellers
| Seller | Health | Rep↓ | Sales | Video | Routing | $ / s | $ / video |
|---|---|---|---|---|---|---|---|
| Venice.ai Proxy 0x1f22…18c9 | healthy | 86.84 | $0.28 | 4 s | #1 | $0.1589 | — |
| 💬 Banter 0x4122…e194 | — | 82.88 | $1.60 | 21 s | #2 | $0.167 | — |
| ▲ Apex Ant 0x73b4…e736 | healthy | 81.71 | $121.04 | 24.3 min | #3 | $0.13 | — |
"Best price" and the seller table are live Antseed catalog data (advertised $/1M tokens — or $ per generated image for unit-billed image models — not settled amounts). Reputation = buyer trust score (0-100, the Antseed SDK's own formula). "Routing" = the SDK's default buyer routing (what the VPR desktop app ships with): a trust ≥ 60 gate on the effective reputation, then cheapest-first among routable sellers; live failover state (per-peer cooldowns) is buyer-side runtime and not included. Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase. "Usage on Antseed" counts only settlements whose buyers share the per-model split on-chain (metadata v2/v3, opt-in), so every usage figure is a lower bound.