Vidu Q4 Preview: Tutorial, Pricing & Latest News

Last checked: October 10, 2026. Vidu Q4 Preview is now available from Vidu and fal.ai. This guide covers the model, a first-generation workflow, API pricing and the release news. InstantVid currently offers Vidu Q3 Pro; this article does not announce a Q4 integration on InstantVid.

What is Vidu Q4 Preview?

Vidu Q4 Preview is ShengShu Technology's latest preview video model. The company announced it on October 7, 2026, with an emphasis on expressive performances and controllable scenes for short dramas, ads and social videos. These are the developer's stated capabilities, rather than results from an InstantVid benchmark. Read the official launch announcement.

The official Q4 model page describes synchronized sound and dialogue, reference voices, more expressive characters, directed camera movement and output options up to 4K. “Preview” remains part of the model name; availability alone does not establish that it is the final release.

Image-to-video or reference-to-video: which should you use?

Workflow	Best starting point	Inputs
Image-to-video	Animate one composition, product photo or character portrait	One starting image and a motion prompt
Reference-to-video	Combine a character, outfit, location or voice	Multiple reference images and optional voice clips

The Vidu reference API accepts up to 15 images and 3 audio references. On fal's reference endpoint, the current image limit is 12, with up to 3 voice clips. Do not assume the two providers have identical limits.

Both Vidu API modes currently accept 3–16 seconds, with 540p, 720p, 1080p, 2K and 4K options. The model's marketing page lists a 1–16 second range for reference generation, but the current API minimum is 3 seconds. Use the chosen provider's current form or API schema when submitting a task. Vidu image API.

Vidu Q4 Preview tutorial: make your first clip

The following is an editorial starter workflow, not a claim that we have tested this prompt with Q4.

1. Open the Q4 image-to-video playground on fal.ai, sign in and check your available balance and quoted price.
2. Upload a clear starting image. Keep the main subject visible, and prepare the composition for the final portrait or landscape format. PNG, JPEG and WebP are supported; the documented limit is 50 MB per image.
3. Describe one main action, the camera movement and the sound you want. Start with 5 seconds at 720p to evaluate motion before paying for a higher resolution.
4. Check the inputs and price, submit the generation and review the result. Inspect the subject's appearance, motion, dialogue and sound before trying a longer clip.
5. Download the finished video and keep your own copy. For Vidu's direct API, documented result links expire after 24 hours.

An original product-video prompt to adapt:

A ceramic coffee cup sits beside a sunlit kitchen window. A hand lifts the cup slowly while steam curls upward. The camera gently pushes closer in one continuous shot. Warm morning light, quiet room ambience and a soft ceramic sound. Keep the cup's shape and decoration consistent.

For reference generation, start with two or three useful images rather than filling every slot. On fal, assign references with tags such as [@reference_image_1] and [@reference_image_2]; a voice reference uses [reference_audio_1]. For example: The character in [@reference_image_1] stands in the room shown in [@reference_image_2], looks toward the window and says "Good morning" using [reference_audio_1]. Use images and voices you have permission to use. fal reference input documentation.

For developers

Vidu's direct API uses the model identifier viduq4-preview, with /ent/v2/img2video and /ent/v2/reference2video creation routes. On fal, use fal-ai/vidu/q4/image-to-video or fal-ai/vidu/q4/reference-to-video.

For fal image generation, the core input fields are image_url, prompt, duration and resolution. Its current image schema does not expose an audio switch; the reference schema does. Keep provider keys on your server and follow each endpoint's own examples. fal image API guide.

Vidu Q4 Preview pricing comparison

Prices below were checked on October 10, 2026. They are provider generation costs, not InstantVid's retail prices or credits. Compare the same duration, resolution and promotion period.

fal: USD pricing and launch promotion

The published promotion is 30% off through November 30, 2026. These rates apply per output second; a five-second example makes the total easier to budget. fal pricing.

Resolution	Regular USD/second	Promotion USD/second	5 seconds, promotion
540p	$0.045	$0.0315	$0.1575
720p	$0.095	$0.0665	$0.3325
1080p	$0.120	$0.0840	$0.4200
2K	$0.190	$0.1330	$0.6650
4K	$0.390	$0.2730	$1.3650

For example, an eight-second 1080p generation works out to $0.672 during the promotion, versus $0.96 at the listed regular rate. This is calculated from the rate card, not a paid test result.

Vidu direct API: provider credits and RMB

The official API price table lists RMB 0.03125 per Vidu credit. The following amounts use its listed credit rates, without applying an account-specific promotion. Values are rounded to four decimal places.

Resolution	Vidu credits/second	Credits for 5 seconds	5 seconds, listed RMB cost
540p	9	45	¥1.4063
720p	19	95	¥2.9688
1080p	24	120	¥3.7500
2K	38	190	¥5.9375
4K	78	390	¥12.1875

At the official regular image-to-video rate, Q4 uses 19 credits/second at 720p versus Q3 Pro's 20: a 5% reduction in credits. Their listed 540p and 1080p rates are the same. Q3 off-peak rates are a separate offer and should not be assumed to apply to Q4.

The launch announcement also mentions a 30% discount, but the official credit table does not show it in these rows. Check your account's quote. The USD and RMB columns use different currencies and pricing conditions; without an exchange rate and matching offers, they do not establish which provider is cheaper.

Latest Vidu Q4 Preview news

• October 7, 2026: ShengShu announced Q4 Preview, emphasizing performances and cinematic generation, including claimed 10-bit color support. This is a manufacturer announcement, not our independent quality assessment.
• Checked October 10, 2026: Vidu and fal publish image-to-video and reference-to-video documentation. Their reference limits differ, so tutorials and integrations should follow the selected endpoint.
• November 30, 2026: The advertised launch discount is scheduled to end. This is a promotion date, not an announced date for a final Q4 release.

The available sources do not establish a guaranteed generation time, an independent model ranking or a final-release timetable. Recheck the linked documentation before a production rollout.

Can I use Vidu Q4 Preview on InstantVid?

Q4 Preview is not currently integrated into InstantVid. The Vidu option available here is Vidu Q3 Pro. You can explore the current video models, start creating with an available model, or read more AI video news.