How to Use Hailuo 2.3 Image-to-Video Pro: Specs, Credits and Prompting

Hailuo 2.3 Image-to-Video Pro is the premium image-animation tier of MiniMax’s Hailuo 2.3 line on VdoBloom: it turns one still image into a 6 or 10 second clip at 768P or 1080P for 25 to 45 credits, it has no text-only mode, and it runs STRICT content filtering on both your image and your prompt. It is the tier you use when the clip is the deliverable, not a test.

Disclosure: VdoBloom is our platform. Competitor details were checked against their own site and we name the competitor as the better pick wherever that is true.

What this model does, honestly

Hailuo 2.3 Image-to-Video Pro does exactly one job: it takes a frame you have already art-directed and puts it in motion. A photograph, a rendered character, a product shot, a still generated elsewhere on the platform. The image is required — it defines the look, the composition, the lighting and the identity of everything on screen. An optional text prompt then directs what happens inside that frame.

That narrow scope is the point. You are not asking the model to invent a world; you are asking it to respect one and animate it. In practice, Pro holds onto the source frame’s texture and colour better than lighter tiers, which is what you want when someone will watch the clip full-screen rather than thumb past it.

It will not generate from text. If you have no image, use a dual-mode model like Hailuo 02, which handles both text-to-video and image-to-video in one place.

If the source image is a photograph of a real person, you must have that person’s consent before animating it.

Real specs and exact credit costs

Two durations, two resolutions, no aspect-ratio control — the output shape follows the source image, so crop before you upload. There is no 10-second 1080P option. Using the Lite plan rate of 300 credits for $15, one credit is $0.05.

Duration & resolution	2.3 Pro credits	Pro cost	2.3 Standard credits	Standard cost
6 seconds, 768P	25	$1.25	15	$0.75
6 seconds, 1080P	45	$2.25	30	$1.50
10 seconds, 768P	45	$2.25	25	$1.25
10 seconds, 1080P	Not supported	—	Not supported	—

Read the Pro column carefully, because the pricing is not linear. Going from 6 seconds to 10 at 768P costs 20 extra credits — an 80 percent increase for 67 percent more footage. Going from 768P to 1080P at 6 seconds costs the same 20 credits. And both upgrades land on the identical 45-credit price, so at 45 credits you are picking one upgrade, never both.

The entry tier at 25 credits ($1.25) is the one to build a habit around. It is the cheapest way to see how this specific model treats your specific image, and everything above it is a deliberate decision rather than a default.

How to prompt it well

On an image-to-video model the prompt is a motion brief, not a scene description. The scene already exists. Describing it again wastes tokens and, worse, invites the model to re-interpret things you had already locked down.

• Name one action. “She turns her head slowly to camera and the hair follows a beat later.” Not three actions in sequence.
• Name one camera move, or none. A slow push-in, or a gentle parallax drift. Holding the frame still is a legitimate and often better choice for product work.
• Say what must not change. Adding a clause such as “preserve facial features, clothing, lighting and background exactly as in the source image” measurably reduces identity drift. VdoBloom’s own effect templates use exactly this pattern.
• Keep it under about 60 words. Long prompts on image-to-video compete with the image instead of guiding it.

On the input side, some images animate far better than others. A single clear subject, decent separation from the background, visible hands and face, and no heavy motion blur already baked in. Busy collages, tight crops that cut the subject at the wrist, and low-resolution screenshots all produce the warping people blame on the model.

A worked example

A jewellery brand has one hero photograph of a necklace on a model and wants a 1080P clip for a paid placement.

Step one: run the image on Hailuo 2.3 Standard at 6 seconds and 768P, 15 credits ($0.75), three times with three different motion briefs — a head turn, a slow push-in on the pendant, and a subtle shoulder shift. That is 45 credits ($2.25) to learn which motion the image actually supports.

Step two: take the winner and re-run the exact same prompt on Pro at 6 seconds and 1080P, 45 credits ($2.25).

Total: 90 credits, $4.50, with the premium render spent only on a motion you have already proven. The expensive alternative is three Pro attempts at 1080P — 135 credits, $6.75 — for the same single usable clip.

When to pick a different model

Choose Standard for anything exploratory, high-volume or disposable: daily social posts, A/B ad variants, motion tests. At 15 credits it is 40 percent cheaper than Pro at the same duration and resolution, and the envelope is identical.

Choose Hailuo 02 when you have no source image, when the shot involves real physics — falling, splashing, collision, momentum — or when the 2.3 tiers’ strict filtering keeps refusing a lawful concept. Hailuo 02 carries the permissive HOT label and starts at 39 credits ($1.95); the flexible-filter overview explains where the platform line actually sits. Swimwear, dance, fitness and fashion concepts that mainstream tools refuse generally belong there, not here.

Choose something else entirely if you need spoken dialogue with matched lip movement, or per-second duration control between 3 and 15 seconds — neither is in this model’s option set. Browse the full lineup in the model directory, and start clips from the image-to-video workspace.

One platform note worth the comparison: MiniMax’s own Hailuo AI product offers Director mode and Subject Reference, which prompt-level control here does not replicate — our VdoBloom vs Hailuo AI page says so directly. What you get on VdoBloom instead is this model in the same credit pool as 105+ others, with one-time packs from $2.49 that never expire against Hailuo’s monthly credit reset. Plans start at $15/month for 300 credits and every paid plan downloads watermark-free with commercial rights — see pricing.

Frequently asked questions

What is the real difference between Hailuo 2.3 Pro and Standard?

The option envelope is identical — both do 6 or 10 seconds at 768P or 1080P. The difference is rendering treatment and price: Pro costs 25, 45 and 45 credits across the three available combinations, Standard costs 15, 30 and 25. Standard is for drafts and volume, Pro for the version an audience sees.

How much does a 1080P Hailuo 2.3 Pro clip cost?

45 credits for 6 seconds at 1080P, which is $2.25 at the Lite plan rate of 300 credits for $15. There is no 1080P option at 10 seconds on either 2.3 tier.

Can Hailuo 2.3 Pro generate video from text alone?

No. A source image is mandatory on both 2.3 tiers — the model is image-to-video only. For text-to-video, Hailuo 02 covers both modes in a single model.

Why was my image rejected?

Hailuo 2.3 Pro is labelled STRICT, meaning the provider screens the uploaded image as well as the prompt text. Mature or sensitive source frames are refused before generation starts. Hailuo 02’s HOT label is more permissive for lawful but edgier concepts; explicit and illegal content is blocked on every model on the platform.

What aspect ratio does the output use?

Whatever your source image uses — there is no aspect-ratio setting. Crop the still to 9:16, 1:1 or 16:9 before uploading and the clip comes back in that shape.