SeedRouter puts image, video, language and audio models behind a single key. Send one request, get the result, pay only for what was delivered. A request that fails costs nothing.
Every card is a model live on the API, with its current starting price. Open one for parameters, a playground and docs.
GPT-6 AstraOpenAICall GPT-6 Astra with the OpenAI SDK. OpenAI's most capable model, with a 1.05M-token context, live per-token pricing and a playground you can try right now.From $8.00 per 1M input tokens
GPT-6 SolOpenAICall GPT-6 Sol with the OpenAI SDK. 1.05M-token context, reasoning effort from none to max, live per-token pricing and a playground you can try right now.From $1.60 per 1M input tokens
Kimi K3Moonshot AICall Kimi K3 with the OpenAI or Anthropic SDK. 1M-token context, always-on reasoning with low, high or max effort, live per-token pricing and a playground.From $2.40 per 1M input tokens
DeepSeek V4.1 FlashDeepSeekCall DeepSeek V4.1 Flash with the OpenAI or Anthropic SDK. 1M-token context, thinking on or off, image input, live peak and off-peak pricing and a playground.From $0.24 per 1M input tokens
Claude Fable 5.1AnthropicCall Claude Fable 5.1 with the Anthropic or OpenAI SDK. Deep reasoning, 1M-token context, cheap cached reads, live pricing and a playground to try it today.From $8.00 per 1M input tokens
Claude Fable 5AnthropicCall Claude Fable 5 with the Anthropic or OpenAI SDK. Deep reasoning, 1M-token context, forced tool use, live per-token pricing and a playground to test it.From $8.00 per 1M input tokens
Claude Opus 5.5AnthropicCall Claude Opus 5.5 with the Anthropic or OpenAI SDK. 1M-token context, adaptive thinking, live per-token pricing and a playground you can try right now.From $3.20 per 1M input tokens
Seedance 2.0ByteDanceGenerate video with Seedance 2.0: text, first and last frames, and image, video or audio references. 480p to 4K, 4–15 seconds. Pay per second, no subscription.From $0.0909 per second
Seedance 2.0 FastByteDanceGenerate video with Seedance 2.0 Fast: the same inputs as Seedance 2.0 at a lower price per second, at 480p or 720p for 4–15 seconds. Pay per second of video.From $0.0548 per second
Seedance 2.0 MiniByteDanceGenerate video with Seedance 2.0 Mini, the lowest price per second in the Seedance 2.0 family: text, frames and references at 480p or 720p, 4–15 seconds.From $0.0182 per second
Seedance 2.5ByteDanceGenerate, edit and extend video with Seedance 2.5: up to 30 seconds at 1080p, with up to 30 image, 10 video and 10 audio references. Pay per second of video.From $0.1329 per second
Nano Banana 2GoogleCreate and edit images with Nano Banana 2 (Gemini 3.1 Flash Image): 512px to 4K, 14 aspect ratios, up to 14 references. Pay per image or per token.From $0.03 per image
Nano Banana ProGoogleCreate and edit images with Nano Banana Pro (Gemini 3 Pro Image): 1K to 4K output, built-in thinking, up to 14 references. Pay per image or per token.From $0.03 per image
Nano Banana 2 LiteGoogleCreate and edit images with Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image), Google's fastest and cheapest image model: 1K output, 14 aspect ratios.From $0.02565 per image
GPT Image 2.5OpenAIGenerate and edit images with GPT Image 2.5 Flare or Sunburst on SeedRouter. Six quality steps up to max, transparent backgrounds, a live playground and pay-as-you-go pricing.From $0.00448 per image
GPT Image 2OpenAICreate and edit images with GPT Image 2 on SeedRouter. Use text prompts, reference images and transparent backgrounds, with a live playground and usage-based pricing.From $0.00425 per image
ByteDance
Google
Claude
Gemini
Alibaba
MiniMax
DeepSeek
Kimi
Mistral
Meta
KuaishouSeedRouter is not a bargain aggregator and not a model showroom. It is the platform layer that makes adding an AI model to a real product a considered choice, not a guess.
See pricingGPT Image 2$0.00425 / imgNano Banana 2$0.03 / img
Seedance 2.5$0.1329 / sClaude Opus 5.5$3.20 / 1MCheck the models SeedRouter supports, what each one can do and what it costs before you write a line. No mystery wrapping and no surprise behaviour: every model has a documented interface.
{
"model": "g",
"prompt": "…",
"api_key": "sk-…"
}Switch model families as your needs change. Move between GPT, Claude and Gemini, or between image and video families, without rewriting your stack. SeedRouter keeps the request shape the same.
→ POST /v1/images/generations
"model": "gpt-image-2"
← 200 task_8f6f…
✓ completed · output on CDN
It starts from model pages and API paths, not demos. Every entry point is designed for an integration that ships and stays up.
Browse the video, image, language and audio models on SeedRouter, and compare them by the capabilities and controls a production workflow needs.
Generate and edit video from text or a first frame, at the resolutions and durations each model supports, with native audio where the model provides it. You get a task id you can poll or receive by webhook.
Explore all video APIsCreate images from text or edit them with natural-language instructions and reference images. Text rendering, consistent characters and product shots are all available on SeedRouter with one request shape.
Explore all image APIsBuild chat, reasoning, code generation and agent workflows with tool calls and structured output. Language models on SeedRouter are served through an OpenAI-compatible endpoint, so existing SDKs work by changing the base URL.
Explore all LLM APIsGenerate full songs, voices and dialogue from a prompt, or separate stems from an existing track. SeedRouter delivers audio as tracks with duration, lyrics and cover art where the model provides them.
Explore all audio APIsPredictable behaviour, clear pricing and real-time control. What SeedRouter gives you out of the box.
Get your API keyOnce
is how often you pay per request
1 task id
per job, from submit to output
SeedRouter bills you when the result is delivered. A failed request costs nothing, and you can retry or switch models with the same key.
1> billing: on delivery
2> failed request: $0
3> retry: same key, any model
Video and long image jobs return a task id. Poll it or get a webhook when the output is ready.
1request → task_id
2poll ← status, progress
3webhook ← output on a CDN link
1
dashboard for every model
Pay per use
pricing model
requests
1
spend
$0.0043
Track usage and see cost per request as it happens, across every model.
1balance, keys, usage
2every request with its price
3top up in USD
Pay per use across every model. See the price before you call.
1pay per request
2no monthly minimum
3price listed per model
Browse the image, video and text models you can call through the SeedRouter API, each with its live starting price. Open one for its parameters, playground and docs.
GPT-6 AstraOpenAIText GenerationFrom$8.00per 1M input tokensGPT-6 SolOpenAIText GenerationFrom$1.60per 1M input tokens
GPT-6 LunaOpenAIText GenerationFrom$0.08per 1M input tokensKimi K3Moonshot AIText GenerationFrom$2.40per 1M input tokensDeepSeek V4.1 FlashDeepSeekText GenerationFrom$0.24per 1M input tokensClaude Fable 5.1AnthropicText GenerationFrom$8.00per 1M input tokensClaude Fable 5AnthropicText GenerationFrom$8.00per 1M input tokensClaude Opus 5.5AnthropicText GenerationFrom$3.20per 1M input tokens
Seedance 2.0ByteDanceText to VideoImage to VideoFrom$0.0909per secondSeedance 2.0 FastByteDanceText to VideoImage to VideoFrom$0.0548per secondSeedance 2.0 MiniByteDanceText to VideoImage to VideoFrom$0.0182per secondSeedance 2.5ByteDanceText to VideoImage to VideoFrom$0.1329per secondNano Banana 2GoogleText to ImageImage to ImageFrom$0.03per imageNano Banana ProGoogleText to ImageImage to ImageFrom$0.03per imageNano Banana 2 LiteGoogleText to ImageImage to ImageFrom$0.02565per imageGPT Image 2.5OpenAIText to ImageImage to ImageFrom$0.00448per imageGPT Image 2OpenAIText to ImageImage to ImageFrom$0.00425per imagePrices are charged per request against your balance. A request that fails is not charged. A model priced in tiers is listed at its standard rate, and a longer request can settle at a higher tier.
One endpoint, compatible with the OpenAI SDK and standard HTTP clients. Move an existing integration to SeedRouter without rewriting it.
See pricingContinue with Googleaccount ready
Create a SeedRouter account in under a minute. Look around before you load any balance.
One key works for every model on SeedRouter. Manage and revoke keys from the dashboard.
base_url="…seedrouter.ai/v1"
Use the OpenAI SDK for language models, or the SeedRouter task API for images, video and audio.
$ curl …/v1/images/generations
← 200 {"id":"task_…"}
Change the base URL and the key. Your first request runs the same day.
One key. Every model. Pay per request. Start building on SeedRouter today.
Everything you need to know before building on SeedRouter.
SeedRouter is one API in front of image, video, language and audio models. You create one key, load a balance and call any model with the same request shape and the same billing. It replaces separate accounts, keys and invoices for each model vendor with a single integration.
Every model on SeedRouter has a listed price: per request, per image, per second of video or per token. You are charged when the output is delivered and nothing for failed requests. There is no subscription and no monthly minimum.
The request returns an error instead of a result, and you are not charged for it. Retry when the model is back, or send the same request to another model on SeedRouter with the same key and the same request shape.
Yes. Language models are served through an OpenAI-compatible endpoint. Point the SDK at SeedRouter, use your SeedRouter key and keep your existing code.
Video and some image and audio models run as tasks. The request returns a task id; poll it or register a webhook, and the output arrives as a CDN link when it is ready.
Image, video, language and audio families, including GPT, Claude, Gemini and GPT Image, plus video and image families such as Seedance, Flux and Qwen Image. The full list with prices is on the models page, and it grows without changes on your side.
No. Top up a balance in USD and spend it across every model on SeedRouter. Unused balance stays in your account.
Yes. SeedRouter is built for products, not demos: per-request billing you can audit, webhooks for long jobs, and no charge for requests that fail.
Guides, comparisons, and deep dives on AI image, video, and model APIs.
Claude Opus 5.5
Anthropic's prompting advice for Claude Fable 5.1 and Opus 5.5: which effort to start at, the system prompt lines that fix common behaviors, and what changed.
SeedRouter
Max output tokens
Context window and max output tokens explained: limits for GPT-6, Claude, DeepSeek V4.1 Flash and Kimi K3, the parameter names, and fixes for cut-off replies.
SeedRouter
Specs
Every DeepSeek V4.1 Flash spec in one place: size, active parameters, the 1M context window, the 393,216-token output cap and its low defaults, images and APIs.
SeedRouter
Alternatives
fal.ai alternatives compared on what matters when you switch: free credits, whether credits expire, how you pay, which models they carry, and how you call them.
SeedRouter