Skip to content
Alpha
  • Models
  • Rankings
  • Ori
Talk to sales

Footer

OpenRouterOpenRouter
© 2026 OpenRouter, Inc

Product

  • Chat
  • Rankings
  • Benchmarks
  • Apps
  • Discover
  • Models
  • Ori
  • Collections
  • Providers
  • Tools
  • Pricing
  • Business
  • Enterprise
  • Labs

Company

  • About
  • Blog
  • Customers
  • Careers
    Hiring
  • Privacy
  • Terms of Service
  • Trust Center
  • Support
  • Works With OR
  • Data
  • Brand

Developer

  • Documentation
  • API Reference
  • Developer Platform
  • Status
  • AI Site Map

Connect

  • Discord
  • GitHub
  • LinkedIn
  • X
  • YouTube
Favicon for bytedance

ByteDance: Seedance 1.5 Pro

Going away November 11, 2026

bytedance/seedance-1-5-pro

BenchmarksCompare

ByteDance's next-generation audio-visual generation model with a 4.5B parameter Dual-Branch Diffusion Transformer architecture. Seedance 1.5 Pro generates video and audio simultaneously in a single unified pass — eliminating the timing issues of sequential audio dubbing. Supports multi-language lip-sync (English, Mandarin, Japanese, Korean, Spanish, and more), cinematic camera control (pan, tilt, zoom, orbit), multi-character dialogue, and character consistency across shots. Produces clips from 4–12 seconds at up to 1080p. The number of tokens is given by (height of output video * width of output video * duration * 24) / 1024

Modalities
Price
from $0.02306/second
Released
Mar 23, 2026
BenchmarksCompare
PlaygroundBenchmarksProvidersPerformanceUptimeAppsActivityDocumentationFAQExplore

Example outputs summary for ByteDance: Seedance 1.5 Pro

  • Text to Video: a slow cinematic push-in on a glowing neon sign that reads "OpenRouter" in the window of a cozy coffee shop on a rainy night, rain streaking down the glass, reflections rippling on wet pavement — generated with ByteDance: Seedance 1.5 Pro

Playground

Benchmarks

Scores on standardized evaluations and this model’s own outputs on shared benchmark prompts. Higher percentages are better — and each rank shows where this model lands among the models evaluated on that source.

Benchmark score summary for ByteDance: Seedance 1.5 Pro (Artificial Analysis)
SourceBenchmarkScore
Artificial AnalysisSeedance 1.5 pro Elo1,256

Benchmark outputs summary for ByteDance: Seedance 1.5 Pro

  • Traffic light: A fixed camera on a single traffic light against a plain grey sky. The light starts red, changes to green, and then changes to amber, in that exact order, each colour holding for a clearly visible moment. No other objects, no camera movement, no cuts. (Tied 18th of 28 on this prompt)
  • Five candles: A fixed straight-on shot of five lit candles in a row on a dark table. The candles go out one at a time from left to right until all five are out. No hand or person visible, no camera movement, no cuts. (Tied 17th of 28 on this prompt)
  • Wrong-way escalator: A fixed shot of an empty upward-moving escalator. A single red rubber ball bounces down the escalator steps from top to bottom, against the direction of the escalator's movement, and exits the frame at the bottom. No people, no cuts. (Tied 17th of 28 on this prompt)
  • Frozen crowd: A busy pedestrian plaza in daylight where every single person stands perfectly still like a statue for the entire clip. Only the flags on the buildings move in the wind. No person moves at all, not even slightly. Fixed camera, no cuts. (Tied 22nd of 28 on this prompt)
  • Sign stays readable: A slow steady dolly toward a shop window sign reading exactly "OPEN 7 DAYS - COFFEE & BAGELS - EST. 2003". The text stays sharp, correctly spelled and unchanged for the whole clip as the camera approaches. No other text in the frame. (27th of 28 on this prompt)
  • Glass of water: A fixed close shot of a full glass of water on a wooden table. A rolling orange bumps into the glass. The glass tips over away from the camera, water spills across the table and spreads outward, and the empty glass rolls slightly and comes to rest. No hands, no cuts. (Tied 15th of 28 on this prompt)
  • Domino run: A fixed side-on shot of exactly eight upright black dominoes in a straight evenly spaced row on a white surface. The leftmost domino tips right and the whole row falls in a single left-to-right chain until all eight are flat. No hands after the start, no cuts. (Tied 23rd of 28 on this prompt)
  • Mirror walk: A fixed shot of a large wall mirror in a bright dance studio. A man in a green t-shirt walks from left to right in front of the mirror. His reflection matches his movement exactly, mirrored, for the entire clip. No cuts. (Tied 4th of 28 on this prompt)
  • Behind the pillar: A fixed camera on a wide walkway with a single thick concrete pillar in the middle. A woman in a yellow raincoat walks at a steady pace from the left edge to the right edge, passing behind the pillar. She is briefly hidden, then reappears on the other side wearing exactly the same clothes, and exits frame right. No cuts. (Tied 1st of 28 on this prompt)
  • Steaming mug: Thin steam rises from the coffee and drifts slowly to the right. Everything else holds still: the mug and its painted design, the table, the background, the lighting and the framing stay exactly as in the reference. Static camera. No cuts. (24th of 27 on this prompt)
  • Walk out: The person from the reference turns and walks out of the frame to the right. Their face, hair and clothing stay exactly as in the reference until they exit. The background does not change. No cuts. (Tied 15th of 26 on this prompt)
  • Continuous shot: A single continuous shot, no cuts, orbiting exactly once, slowly, around a stone statue of a horse in an empty courtyard. The statue stays centred the whole time and the camera ends where it started. Nothing else enters the frame. (Tied 25th of 28 on this prompt)

Providers

This model is hosted by one provider. OpenRouter forwards every request to it directly — no routing decisions to make.

Performance

Throughput is how fast the model writes (tokens per second — higher is better). Latency is total round-trip time (lower is better). TTFT is time-to-first-token — how long before you see anything appear (lower is better).

Uptime

Uptime is the percentage of the past 3 days that at least one provider was responding to requests. Availability is the percentage of time that inference was successfully served. OpenRouter continuously monitors and uses the next-best provider when one returns an error.

Apps

Public apps that send the most traffic to this model. Good signal for what real production workloads look like — and a hint at which use cases this model is best suited for.

Activity

Token volume and request traffic to this model over time.

API

Drop-in code to call this model. OpenRouter's API is OpenAI-compatible — most SDKs work by just swapping the base URL. The only thing that changes between models is the model slug below.

Not currently available through EU in-region routing. In-region routing docs

Documentation

Guides, cookbooks, blog posts, and tools for building with ByteDance: Seedance 1.5 Pro on OpenRouter.

  • Best Image-to-Video AI Models in 2026: Cost, Resolution, and ControlBlog post

Explore more models

AI Models with Vision: Multimodal LLMs for Image UnderstandingCollectionVideo Generation ModelsCollectionVideo Model RankingsRanking
End-to-end latency
$2.40$0.05184$1.20$0.0259256.98s
96.55%
End-to-end latency
56.98s

P50, best provider

Uptime (3d)The model was reachable. Request routed to a provider.
100.00%
Availability (3d)The model returned inference from any provider. Errors and empty responses count against it.
99.15%

Availability over the last 3 days

Last 72 hours
Availability 99.15%
3 Days Ago2 Days AgoYesterdayNow

Availability over the last 24 hours

OpenRouter Availability
99.30%

When an error occurs in an upstream provider, we can recover by routing to another healthy provider, if your request filters allow it. You can access per-provider uptime data programmatically through the Endpoints API. Learn more about our load balancing and provider routing options.

1.
Favicon for https://komrz4.mykomrz.com/
komrz4
new
0tokens
2.
Favicon for https://shinjibot.com/
ShinjiBot
new
0tokens
3.
Favicon for https://github.com/novastrike
Nova Strike asset pipeline
new
0tokens
4.
Favicon for https://local.studio/
Writing Studio Research
new
0tokens
5.
Favicon for https://crelivo.local/prompt-library-ai
Prompt Library AI
new
0tokens

No chart data available

Frequently asked questions

ByteDance's next-generation audio-visual generation model with a 4.5B parameter Dual-Branch Diffusion Transformer architecture. Seedance 1.5 Pro generates video and audio simultaneously in a single unified pass — eliminating the timing issues of sequential audio dubbing.

The Seedance 1.5 Pro documentation on OpenRouter includes the Best Image-to-Video AI Models in 2026: Cost, Resolution, and Control blog post. The quickstart covers API keys and your first request, and the API reference lists every parameter.

Seedance 1.5 Pro costs from $0.02306/second at 480p to $0.4666/second at 4K for Video (with audio), from $0.01153/second at 480p to $0.2333/second at 4K for Video (no audio), $2.40/M tokens for Video Tokens (with audio), and $1.20/M tokens for Video Tokens (no audio).

Seedance 1.5 Pro supports 4–12 second clips, can be steered with a first frame and last frame image, and produces a matching audio track.

Seedance 1.5 Pro accepts text and images as input and returns video.

Seedance 2.0 Mini, Seedance 2.5, Seedance 2.0 and 1 more are other video models from ByteDance.

Seedance 1.5 Pro was released on March 23, 2026.

More models from bytedance

Seedance 2.0 Mini

Seedance 2.0 Mini is a video generation model from ByteDance. It supports text-to-video, image-to-video with first and last frame control, and multimodal reference-to-video with image, video, and audio inputs. It supports 480p and 720p output for 4-15 second videos. The number of tokens is given by (height of output video * width of output video * duration * 24) / 1024

Videofrom $0.03363/second
Seedance 2.5

Seedance 2.5 is a video generation model from ByteDance. It is suited for long-form storytelling, multimodal reference-based generation, video editing, and video extension. It supports first-frame and first-and-last-frame control, up to 50 image, video, and audio reference assets, optional generated audio, and multilingual audiovisual generation.

Videofrom $0.1028/second
Seedance 2.0 Fast

Seedance 2.0 Fast is a video generation model from ByteDance. It supports text-to-video, image-to-video with first and last frame control, and multimodal reference-to-video. It prioritizes generation speed and lower cost over maximum output quality. The number of tokens is given by (height of output video * width of output video * duration * 24) / 1024

Videofrom $0.04035/second
Seedance 2.0

Seedance 2.0 is a video generation model from ByteDance. It supports text-to-video, image-to-video with first and last frame control, and multimodal reference-to-video. It is particularly strong at preserving character consistency, visual style, and camera movement from reference material. The number of tokens is given by (height of output video * width of output video * duration * 24) / 1024

Videofrom $0.06726/second
Seed OSS 36B Instruct

Seed-OSS-36B-Instruct is a 36B-parameter instruction-tuned reasoning language model from ByteDance’s Seed team, released under Apache-2.0. The model is optimized for general instruction following with strong performance in reasoning, mathematics, coding, tool use/agentic workflows, and multilingual tasks, and is intended for international (i18n) use cases. It is not currently possible to control the reasoning effort.

Text131K context
UI-TARS 7B

UI-TARS-1.5 is a multimodal vision-language agent optimized for GUI-based environments, including desktop interfaces, web browsers, mobile systems, and games. Built by ByteDance, it builds upon the UI-TARS framework with reinforcement learning-based reasoning, enabling robust action planning and execution across virtual interfaces.

This model achieves state-of-the-art results on a range of interactive and grounding benchmarks, including OSworld, WebVoyager, AndroidWorld, and ScreenSpot. It also demonstrates perfect task completion across diverse Poki games and outperforms prior models in Minecraft agent tasks. UI-TARS-1.5 supports thought decomposition during inference and shows strong scaling across variants, with the 1.5 version notably exceeding the performance of earlier 72B and 7B checkpoints.

Text128K context$0.10 / $0.20