An OpenRouter alternative for video and image generation

Competitor prices checked October 22, 2026 from OpenRouter's public pricing; they change often.

Prices as of October 22, 2026 (price list v5).

How OpenRouter's video and image APIs work

• Video is asynchronous. POST /api/v1/videos starts a job; you poll it and download the result. A callback_url webhook is supported. GET /api/v1/videos/models lists the video models with their pricing_skus; the cost of a job appears in usage.cost once it completes.
• Pricing. OpenRouter says it adds no markup on inference pricing and charges a fee when you buy credits: 5.5% on the Standard plan, 8% on Business, with a separate fee for crypto. Bring-your-own-key use is free up to $25k a month, then 5%.
• Refunds. Unused credits are refundable within 24 hours; platform fees aren't refundable, and crypto purchases never are.
• Failed jobs. OpenRouter's "zero completion insurance" covers text, TTS and image requests with no usable output. For video, the docs list failed, cancelled and expired statuses but don't state whether those jobs are billed (checked Oct 22, 2026).

Which models overlap?

Per-second prices, model by model

Fees and billing: credit-purchase fee vs prepaid wallet, quotes, failed jobs

When to stay on OpenRouter

• You also call LLMs. OpenRouter is built around text models. Makernice has no LLM text API.
• You bring your own provider keys. OpenRouter's BYOK is free up to $25k a month.
• You rely on the OpenAI SDK. OpenRouter's ecosystem is built around OpenAI-style clients (OpenRouter says its API is OpenAI-compatible; openrouter.ai model pages, checked Oct 22, 2026). Makernice is not compatible with the OpenAI SDK: it is its own async REST API (quote, create, poll, cancel), so moving means rewriting the calls, not swapping a base URL.
• You need a model we don't have. OpenRouter's video list (30 models on October 22, 2026) includes models Makernice doesn't offer, and its overall catalog is far larger.
• Your mix is mostly Seedance at 720p, Happy Horse 1.1 or Qwen image-to-image with one or two input images. Those cost less there, even after the 5.5% credit fee.

Calling the same model on Makernice

import os, time, uuid, requests

API = "https://api.makernice.com/v1"
H = {"Authorization": f"Bearer {os.environ['MAKERNICE_API_KEY']}"}

job = {"model": "wan-3.0", "mode": "flf",
       "prompt": "The flower opens in time-lapse, morning light.",
       "image_url": "https://example.com/bud.jpg", "last_image_url": "https://example.com/bloom.jpg",
       "duration": 6, "resolution": "720p", "generate_audio": False}

quote = requests.post(f"{API}/generations/quote", json=job, headers=H).json()
print(f"${int(quote['amount']) / 1e6:.2f}")          # $0.47 at v5, locked for 12 minutes
gen = requests.post(f"{API}/generations", json={"quote_id": quote["id"], "input": job},
                    headers={**H, "Idempotency-Key": str(uuid.uuid4())}).json()

while gen["status"] not in ("succeeded", "failed", "canceled"):
    time.sleep(5)
    gen = requests.get(f"{API}/generations/{gen['id']}", headers=H).json()

About Makernice

Makernice is an AI video and image generation API: one API key and one prepaid balance for 16 models from ByteDance, Kuaishou, MiniMax and Alibaba. Every run starts with a quote that locks the price for 12 minutes, through Makernice's own async REST API (quote, create, poll, cancel).

Frequently asked questions

Related

• All 16 models · Wan 3.0 · Wan 3.0 Prime · Happy Horse 1.1 · Seedream 5.0 Pro · Qwen3 Image
• Guides: Image generation pricing · Image-to-video and first/last frame
• Billing: Pricing · No charge for failed generations
• Other alternatives: Replicate alternative · WaveSpeed alternative · Atlas Cloud alternative · Higgsfield API alternative
• Cost: AI video cost calculator

Product and company names on this page are trademarks of their respective owners. Makernice is not affiliated with, endorsed by or sponsored by them.