Three matched standard routes
Hailuo 2.3 API pricing
Compare one exact Hailuo 2.3 standard text-to-video request at 768p and six seconds without mixing fast, pro, image-to-video, or 1080p offers.
Hailuo 2.3 API pricing
Hailuo 2.3 standard 768p six-second price
- Reviewed comparison scope
- Hailuo 2.3 · standard · text to video · 768p · 6 seconds · no audio · global · USD per request
- Publication gate
- At least 3 unique providers with accepted, fresh, reviewed evidence
| Price rank | Provider | Route and tier | Published meter | Evidence |
|---|---|---|---|---|
| 1Tie | fal.ai | aggregator · standard | $0.28 / source requestfresh | First-party sourceChecked Aug 23, 2026Fresh until Aug 30, 2026 |
| 1Tie | Novita AI | aggregator · standard | $0.28 / source requestfresh | First-party sourceChecked Aug 23, 2026Fresh until Aug 30, 2026 |
| 1Tie | Replicate | marketplace · standard | $0.28 / source requestfresh | First-party sourceChecked Aug 23, 2026Fresh until Aug 30, 2026 |
Price meter only: Fast, pro, image-to-video, ten-second, and 1080p SKUs are separate. An older WaveSpeedAI blog says $0.23 while its current route says $0.28, so WaveSpeedAI is excluded until that conflict is cleanly retired or explained. Equal prices share a price rank; provider name and offer ID determine only their stable display order.
Configuration-specific buying guide
How to use this Hailuo 2.3 API pricing result
The Hailuo 2.3 family contains standard, fast, pro, text-to-video, and image-to-video routes. This page does not blend them. It compares the standard text-to-video product at 768p for exactly six seconds. fal, Replicate, and Novita each publish twenty-eight cents for that SKU, so the result is a three-way price tie rather than an invented winner.
WaveSpeedAI is intentionally absent from the eligible sample. Its current model surface states twenty-eight cents for the standard text-to-video route, but an older provider blog still states twenty-three cents for the same six-second 768p description. The newer page may be correct, yet APIDir will not select whichever number produces a more attractive ranking while both provider-owned statements remain discoverable.
A tied meter does not remove provider differences. Request schema, concurrency, queue time, retention, content policy, webhook behavior, support, and failure billing still require testing. Use the price result to confirm the baseline, then choose on operational fit and measured usable-output cost.
What changes the real cost
- Standard, fast, and pro are different SKUs with distinct price and quality promises.
- Six versus ten seconds and 768p versus 1080p change the fixed price per completed video.
- Text-to-video cannot inherit the price of an image-to-video route even when the duration matches.
- Retries, moderated jobs, output storage, delivery, and human review can outweigh a small route-level difference.
Checks before production
- Pin the text-to-video standard endpoint and reject a generic Hailuo family key.
- Validate duration and resolution before submission so a ten-second or 1080p request cannot use the six-second estimate.
- Use a bounded polling schedule and expose terminal provider errors instead of converting them into empty success responses.
- Reconcile completed job configuration and charged amount against the provider's current pricing record.
Workloads this result can inform
Six-second creative tests
The tied table is a useful baseline. Run identical prompts and measure acceptance rate, queue time, and manual rejection before selecting a channel.
Longer social clips
Move to the documented ten-second SKU. Do not multiply the six-second request price because providers publish a separate fixed total.
Image-led animation
Use the image-to-video price and schema. The reference image changes the task even when the provider markets both routes under Hailuo 2.3.
Public methodology
How an offer enters the ranking
- The observation passed APIDir's accepted evidence and publication-permission gates and retains an official first-party source.
- Its offer identity, provider channel, route, and tier are explicitly included after configuration review.
- Its freshness label is fresh and its fresh-until time has not passed.
- The model, billing unit, region, and every reviewed comparison field match the group definition.
- At least three unique providers qualify; otherwise APIDir publishes no order, cheapest claim, or spread.
- Prices sort low to high. Equal prices share a rank; provider name, then offer ID, controls only stable display order.
Questions developers ask
When does APIDir publish a cheapest provider?
Only when at least three unique providers have accepted, publication-permitted, fresh observations for the same model and billing meter, with every route and material configuration field explicitly reviewed.
What happens to stale or quarantined prices?
They are excluded from ranking order. Historical evidence may remain auditable elsewhere, but it cannot support a current winner or savings claim.
Does the first row mean the provider is best overall?
No. The order covers the displayed price meter only and does not score quality, speed, uptime, limits, support, legal terms, or total production cost.
APIDir Delta · controlled rollout
Join the source-linked pricing update waitlist.
Confirm your address to join the waitlist. Editorial delivery starts only after sender, suppression, and evidence checks are active.