WaveSpeed AI API pricing per model: Seedance, Nano Banana and GPT-6 rates compared
WaveSpeed AI API pricing per model for Seedance, Nano Banana, GPT-6 and Claude, its $1 free credit and account levels, and where the same models cost less.
Read as MarkdownWaveSpeed AI API pricing is pay-per-use: you add credits, each model page lists its own price per second of video, per image or per million tokens, and there are no monthly fees. Several models also showed a 10% discount on October 6, 2026. This guide reads those per-model prices for Seedance, Nano Banana and the GPT-6 and Claude language models, and compares them with SeedRouter's rates for the same models.
SeedRouter is the cheaper option compared here, and it is our product. WaveSpeed has a far larger catalog and was the lower price on some models, including low-quality GPT Image renders and DeepSeek V4.1 Flash at peak hours, so this guide only tables the models where SeedRouter cost less.
Third-party prices in this guide were checked on October 6, 2026.
How does WaveSpeed AI pricing work?
WaveSpeed bills each generation from a prepaid credit balance at the price on the model's page. Video models are priced per second by resolution, image models per image by resolution, and language models per million input and output tokens. WaveSpeed accepts cards, PayPal, Google Pay, Apple Pay, WeChat Pay and Alipay, and offers enterprise plans with volume discounts.
Account levels depend on your top-ups: Bronze by default, Silver after any successful top-up below $1,000, Gold after a single top-up of $1,000 to $4,999, and Ultra after a single top-up of $5,000 or more.
Does WaveSpeed AI give free credits?
Yes, a small amount. WaveSpeed says eligible new accounts receive $1 in free credits with no credit card, and that some premium models may not be available with trial credits. Every new SeedRouter account also starts with a small free balance, and SeedRouter credits never expire.
What does WaveSpeed AI charge for Seedance?
WaveSpeed lists Seedance per 5 seconds by resolution; the figures below are its per-second rates after the 10% discount shown on the Seedance 2.5 and 2.0 pages, which is the most favourable reading for WaveSpeed. Seedance 2.0 Fast and Mini showed no discount. SeedRouter's figures are the per-second prices for 16:9 video on its model pages.
| Model and resolution | WaveSpeed AI | SeedRouter | Difference |
|---|---|---|---|
| Seedance 2.5, 480p / 720p / 1080p | $0.162 / $0.324 / $0.81 per s | $0.117 / $0.263 / $0.647 per s | −28% / −19% / −20% |
| Seedance 2.0, 480p / 720p | $0.108 / $0.216 per s | $0.080 / $0.172 per s | −26% / −20% |
| Seedance 2.0, 1080p / 4K | $0.54 / $1.08 per s | $0.426 / $0.883 per s | −21% / −18% |
| Seedance 2.0 Fast, 480p / 720p | $0.10 / $0.20 per s | $0.055 / $0.118 per s | −45% / −41% |
| Seedance 2.0 Mini, 480p / 720p | $0.06 / $0.12 per s | $0.018 / $0.039 per s | −70% / −68% |
A 10-second Seedance 2.5 clip at 1080p comes to $8.10 at WaveSpeed's discounted rate. On SeedRouter it costs about a fifth less, and Seedance 2.0 Mini drafts cost less than a third of WaveSpeed's rate.
What does WaveSpeed AI charge for Nano Banana?
Both Nano Banana models showed the 10% discount on WaveSpeed; the figures below include it. SeedRouter's standard model IDs for Nano Banana Pro and Nano Banana 2 charge one flat price per delivered image at any size.
| Model and size | WaveSpeed AI | SeedRouter | Difference |
|---|---|---|---|
| Nano Banana Pro, 1K / 2K | $0.126 | $0.03 | −76% |
| Nano Banana Pro, 4K | $0.216 | $0.03 | −86% |
| Nano Banana 2, 0.5K / 1K | $0.0405 / $0.063 | $0.03 | −26% / −52% |
| Nano Banana 2, 2K / 4K | $0.0945 / $0.126 | $0.03 | −68% / −76% |
What does WaveSpeed AI charge for GPT-6 and Claude?
WaveSpeed lists language models per million tokens at their list prices. SeedRouter's GPT-6 figures are its rates for prompts up to 272K tokens.
| Model | WaveSpeed AI input / output | SeedRouter input / output | Difference |
|---|---|---|---|
| GPT-6 Astra | $10 / $50 | $8 / $40 | −20% / −20% |
| GPT-6 Sol | $2 / $10 | $1.60 / $8 | −20% / −20% |
| GPT-6 Luna | $0.10 / $0.50 | $0.08 / $0.40 | −20% / −20% |
| Claude Opus 5.5 | $4 / $20 | $3.20 / $16 | −20% / −20% |
| Claude Fable 5.1 | $10 / $50 | $8 / $40 | −20% / −20% |
| Kimi K3 | $2.648 / $13.283 | $2.40 / $12 | −9% / −10% |
GPT-6 prompts above 272K tokens move to a higher long-context rate on SeedRouter, so compare both for very long prompts. See GPT-6 Sol and Claude Opus 5.5 for the full rate cards, including cache pricing.
Frequently asked questions
Is WaveSpeed AI free?
Not beyond its trial credit. WaveSpeed is pay-per-use with no subscription, and eligible new accounts get $1 in free credits to try it, which may not cover premium models.
Is WaveSpeed AI legit?
WaveSpeed is an operating API platform with public per-model prices, documentation, an MCP server and enterprise plans. As with any API, run a small paid test with your own prompts before moving production traffic.
How stable is WaveSpeed?
We cannot measure WaveSpeed's reliability from this guide. Its enterprise plans advertise performance SLAs, which the pay-per-use tier does not include. If you need a second source for Seedance or Nano Banana, SeedRouter runs both as tasks that you poll until they finish; see the task docs.
Is WaveSpeed cheaper than Higgsfield or Kie?
Not on Seedance 2.5 at 1080p. WaveSpeed's discounted $0.81 per second was above Higgsfield's API price of $0.7961 and Kie's $0.790, both checked a day earlier, so it was the highest of the three on that tier. On the Seedance and Nano Banana tiers above, SeedRouter listed lower prices than WaveSpeed; our Higgsfield API pricing and Kie API pricing guides compare those two platforms with SeedRouter the same way.
Choosing between WaveSpeed AI API pricing and SeedRouter
WaveSpeed is a strong choice for its wide catalog and the models where it was cheaper on October 6, 2026. For Seedance 2.5, 2.0, Fast and Mini, Nano Banana Pro and Nano Banana 2, and the GPT-6, Claude and Kimi K3 models above, SeedRouter cost 9% to 86% less at the tiers compared. Run the same prompts on both and keep each model where it costs least for you.



