REVIEW LAB · DESK REVIEW · 15 SERVICES

The new
video
machines.

Kling 4 just dropped, Sora is gone, and three Chinese labs are crowding Google at the top. Fifteen AI video services checked this week: how long, how sharp, whether they talk, and what they cost.

By cc, PointCast Review Lab
· 9 minute read

See all fifteen ↓
KLING 4
DROPPED
SEPT 28 ↘A pixel-dithered specimen sheet in black, cobalt and yellow: a vintage movie camera on a tripod, a blank clapperboard, a film strip with a sun, a wave and a palm leaf, a play button, a timeline with diamond keyframes, a stereo waveform between two speakers, an hourglass, a server rack, a magic wand over a monitor, a price tag, a film reel and a red lightning bolt.
Illustration: Astra (Codex) for PointCast, after Mike Hoydich’s specimen sheets

A year ago, AI video meant four wobbly seconds. This month it means thirty seconds, with sound, in one pass.

Kling announced 4.0 on September 28. So far only a lighter Flash version (720p, up to 20 seconds) is open, and only to Ultra yearly members; the full model (30 seconds, up to 4K) is promised for October. The field has settled on 15-to-30-second clips with sound made in the same pass. Alibaba’s Wan 3.0, MiniMax’s H3 and ByteDance’s Seedance now crowd Google’s Gemini Omni Flash at the top of the blind-vote charts. OpenAI is out: the Sora app closed April 26 and its API shut down September 24.

Desk review from makers’ announcements, docs and pricing pages, checked September 29, 2026, plus news coverage where a maker publishes no page. Rankings are the Artificial Analysis Video Arena blind-vote leaderboards as seen September 29. PointCast has not run a side-by-side test of these models, so there is no star rating. Where a number comes only from resellers or aggregators, the card says so.

WHO’S WINNING THE BLIND VOTE · SEP 29, 2026

The chart, this week.

  1. 01Gemini Omni FlashGoogle1233
  2. 02Wan 3.0Alibaba1229
  3. 03MiniMax H3 Max (tuned by fal)fal / MiniMax1227
  4. 04MiniMax H3MiniMax1220
  5. 05Seedance 2.0 (720p)ByteDance1210
  6. 06MAGI-2 PreviewSand.ai1156

Artificial Analysis Text-to-Video Arena, with audio, Elo from blind head-to-head votes. Kling 4.0 and Seedance 2.5 are too new to be ranked. Without audio, Wan 3.0 leads (1335), just ahead of Omni Flash (1332); Veo 3.1 is sixth and Kling 3.0 seventh. See the live chart ↗

FOUR KINDS OF VIDEO MACHINE

What do you want
to make?

  1. 05 PICKSJust droppedThe August–September launches, and the one everyone’s asking about.
  2. 04 PICKSStudio workbenchesEditing, control and a timeline, not just a prompt box.
  3. 04 PICKSApps & model shopsOne subscription, many models, quick social clips.
  4. 02 PICKSRun it yourselfOpen weights on your own GPU. No meter running.
Compare all ↓

Showing all 15 services.

01● NEW

Kling 4.0

Length
30 s (Flash: 20 s)
Picture
Up to 4K promised (Flash: 720p)
Sound
Yes, stereo
Latest
Sep 28, 2026

Just dropped · The one that just dropped

Kling 4.0

From ~$10/mo*Kuaishou · Kling AI

Early access: 4.0 Flash for Ultra yearly members only. Full model promised for October.

Kling 4.0 is built for directing a scene, not rolling the dice on a prompt. The full model promises 30-second clips at up to 4K with stereo sound, up to 10 keyframes to pin the story beats, and up to 15 references (images, clips, characters, voices) to keep a face or a product the same from shot to shot. You can also edit an existing clip without regenerating it. What you can actually use today is smaller: the Flash version, 720p and up to 20 seconds, open since September 28 only to Ultra yearly subscribers.

What’s new
Announced September 28, 2026: single clips go from 15 to 30 seconds, with keyframes, references and native stereo audio.
Best for
Long, multi-shot scenes you direct beat by beat
Where
kling.ai web and app. Kling 3.x is also on the API and inside Adobe Firefly, Runway and Higgsfield.

The catch. Almost nobody can use it yet. Flash is 720p and has no keyframes, and it’s gated to the top yearly plan until October. One news report says the full model launched September 28. It didn’t. Source ↗

The one to watch. Wait for October unless you already pay for Ultra.

Go to Kling →

Direct link · no commission · * Kling’s Standard plan, as listed by resellers; Ultra runs about $180 a month. Kling hasn’t priced 4.0 yet. Check kling.ai. · Kling’s announcement ↗

02● NEW

Gemini Omni Flash

Length
3–10 s per clip (extends to ~40 s)
Picture
Up to 4K
Sound
Yes
Latest
Aug 27, 2026

Just dropped · Top of the charts

Gemini Omni Flash

~$0.10/s API*Google DeepMind

Live in the Gemini app, Flow, YouTube Shorts and the Gemini API

Google’s new default isn’t Veo. Omni Flash is a Gemini model that makes video the way you’d have a conversation: generate a clip, then ask for changes in plain words. It sits first on the Artificial Analysis blind-vote chart for video with sound, and because it lives in the Gemini app and YouTube, it’s the one most people already have. Base clips are short, but you can extend them.

What’s new
Announced at I/O in May, generally available as Omni 1.1 Flash in late August, and now Google’s recommended default over Veo.
Best for
Generate, then edit by talking to it
Where
Gemini app, Google Flow, YouTube Create, and the Gemini API, where Google now recommends it as the default video model.

The catch. Short base clips. Editing speech is restricted, and complex motion and on-screen text are still weak spots. The late-August release date and the 40-second extension come from secondary reports. Source ↗

The best place to start this week, and probably already in your phone.

Open Gemini →

Direct link · no commission · * Gemini API: $17.50 per million video output tokens at 5,792 tokens a second of 720p video, or about ten cents a second. Also included in Gemini app plans. · Gemini API pricing ↗

03● NEW

Wan 3.0

Length
30 s
Picture
1080p
Sound
Yes
Latest
Aug 24, 2026

Just dropped · Best picture without sound

Wan 3.0

$0.05–0.20/s API*Alibaba · Tongyi Lab

Live on Alibaba Cloud and the Qwen Cloud API, by application

Wan 3.0 is first on the Artificial Analysis chart for video without sound and second with it. Its odd strength is input: hand it a PDF, a slide deck, a spreadsheet or a web address and it will make a video about it. Clips run 30 seconds in one pass.

What’s new
Public beta August 6, launched August 24: single clips doubled to 30 seconds, and documents and URLs are accepted as prompts.
Best for
Turning documents and web pages into video
Where
Alibaba Cloud Model Studio and the Qwen Cloud API, after an application. A members-only consumer site at wan.video is planned.

The catch. The weights are closed now, which ends the open Wan line people ran at home. API access means applying first. We found no official Alibaba page, so these details come from several news reports. Source ↗

Top-tier picture at a low per-second price, if Alibaba lets you in.

Go to Model Studio →

Direct link · no commission · * API: $0.05 a second at 480p, $0.10 at 720p, $0.20 at 1080p. · TechNode on Wan 3.0 ↗

04● NEW

Seedance 2.5

Length
30 s (3-minute mode in beta)
Picture
Up to 4K
Sound
Yes
Latest
Jul 31, 2026

Just dropped · Longest scenes you can make today

Seedance 2.5

From $15/mo*ByteDance · Dreamina

Live in Dreamina (CapCut) worldwide; hosted on Pika, Higgsfield and others

Seedance 2.5 does now what Kling 4 promises for October: 30-second continuous clips, up to 50 reference images and clips to keep characters the same, and local edits that fix one detail without regenerating the whole thing. A beta long mode stretches to three minutes. Its older sibling, Seedance 2.0, is fifth on the with-sound chart; 2.5 is too new to be ranked.

What’s new
Launched July 31: 30-second clips, a beta 3-minute mode, 50 references and local editing.
Best for
Keeping a character consistent across a long scene
Where
Dreamina on the web and in the CapCut family of apps. In China it’s in Jimeng and Doubao.

The catch. The three-minute mode is beta. Plans and prices differ by country, and it’s a ByteDance product, with the data questions that brings. Source ↗

The long-scene pick you can actually use this week.

Open Dreamina →

Direct link · no commission · * Dreamina Basic; the first month was on promotion for $1.50 when we checked. Pricing varies by region. · Dreamina’s launch release ↗

05● NEW

MiniMax H3 (Hailuo 3)

Length
15 s
Picture
2K
Sound
Yes, stereo with lip-sync
Latest
Jul 31, 2026

Just dropped · Cheap 2K with a voice

MiniMax H3 (Hailuo 3)

Price not published*MiniMax

Live in the Hailuo app, on the MiniMax API, and as open weights

H3 takes text, images, video and audio in one prompt and returns 2K video with stereo sound, including dialogue that matches the lips. It’s fourth on the with-sound chart, and a version tuned by fal sits third. Unusually for a model this good, you can download it.

What’s new
Shown at WAIC on July 17 and released July 31 with open weights.
Best for
Talking-head clips and dialogue with sound
Where
Hailuo web and app, the MiniMax API, open weights on Hugging Face (with ComfyUI support), and partners including Luma, Pika and fal.

The catch. Clips top out at 15 seconds. Secondary sources say the open license allows commercial use only for companies under $20 million in revenue. Source ↗

Near the top of the chart, and you can download it.

Open Hailuo →

Direct link · no commission · * MiniMax says 2K output costs under a third of mainstream models but lists no dollar price. The weights are free to download. · MiniMax’s H3 post ↗

06

Runway Gen-4.5

Length
Up to a minute, multi-shot
Picture
1080p
Sound
Yes
Latest
Dec 1, 2025

Studio workbenches · The editor’s workbench

Runway Gen-4.5

From $12/mo*Runway

Live on the web, the app and the API

Runway isn’t the top-ranked model anymore, and it hasn’t shipped a new frontier model since Gen-4.5 in December. What it has is the best editing room: Aleph 2.0 edits existing footage, and since July a Media Router picks between Runway’s models and other companies’ by quality, speed or cost. It’s where you finish a video, not just where you generate one.

What’s new
Aleph 2.0 editing in May; a Media Router that sends each job to Runway’s or a third party’s model, July 23.
Best for
Editing and finishing, with many models in one place
Where
runway.com on the web and app, the Runway API, and inside Adobe Firefly.

The catch. No new generation model in ten months. The API price comes from a reseller. Source ↗

Still the best workbench, even though its own model trails.

Go to Runway →

Direct link · no commission · * Standard plan billed annually: 625 credits, or about 52 seconds of Gen-4.5. Resellers list the API at about $0.12 a second. · Runway pricing ↗

07

Luma Ray3.2

Length
20 s
Picture
1080p, 16-bit HDR / EXR
Sound
No (sound comes from ElevenLabs tools inside Luma)
Latest
Jun 2026

Studio workbenches · For the color grader

Luma Ray3.2

From $25/mo*Luma AI

Live in Luma’s app and API

Luma is still the only one handing out real HDR: 16-bit, exported as EXR files that drop into a professional color pipeline. Ray3.2 adds frame-level direction with up to 16 keyframes per clip. If your video ends up in DaVinci Resolve, this is the one built for you.

What’s new
Ray3.2 adds 16 keyframes per clip and keeps HDR/EXR output. The old Dream Machine and Ray2 are retired.
Best for
HDR footage for a real color grade
Where
lumalabs.ai, the Luma API, and inside Adobe Firefly.

The catch. The model doesn’t make sound; Luma adds it with built-in ElevenLabs tools. Not near the top of the blind-vote charts. Source ↗

The pro’s pick when the file matters more than the prompt.

Go to Luma →

Direct link · no commission · * Plus plan billed annually ($30 monthly): 10,000 credits, or about 100 seconds of 1080p Ray3.2. · Luma pricing ↗

08

Veo 3.1

Length
8 s (extends to 148 s)
Picture
Up to 4K (Lite: 1080p)
Sound
Yes
Latest
Oct 15, 2025

Studio workbenches · Extend and frame-match

Veo 3.1

$0.05–0.40/s API*Google DeepMind

Live on the Gemini API, Vertex AI, Flow and the Gemini app

There’s no Veo 4. Google’s docs now point new work at Omni Flash and keep Veo 3.1 for two jobs it does best: extending a clip scene by scene, and matching a first and last frame you supply. The cheap Lite tier, added this spring, is the lowest per-second price from a major lab.

What’s new
No new model. The cheaper Veo 3.1 Lite arrived in spring 2026.
Best for
Long extensions and exact first/last frames
Where
Gemini API, Vertex AI, Google Flow, the Gemini app, and inside Adobe Firefly.

The catch. Eight-second base clips, and Google now steers new users to Omni Flash. Source ↗

The dependable tool for stitching a longer piece, at five cents a second.

See Veo →

Direct link · no commission · * API: Veo 3.1 Lite $0.05 a second at 720p, Fast $0.10, Standard $0.40. · Gemini API pricing ↗

09

Adobe Firefly

Length
Depends on the model
Picture
Depends on the model
Sound
Depends on the model
Latest
Apr 15, 2026

Studio workbenches · Inside Premiere

Adobe Firefly

Credits; price not checked*Adobe

Live in Firefly and across Creative Cloud

Adobe’s own video model isn’t the draw. The draw is that Firefly now runs more than 30 other companies’ models, including Kling 3.0, Veo 3.1, Runway Gen-4.5 and Luma Ray3.14, from inside the apps editors already use, plus Adobe’s commercially safe model for work where the rights have to be clean.

What’s new
April 2026: Kling 3.0 and 3.0 Omni joined 30-plus partner models, with a new AI assistant and Quick Cut auto-editing.
Best for
Commercially safe footage without leaving Creative Cloud
Where
Firefly on the web and app, Premiere, Photoshop, Express and Firefly Boards.

The catch. Adobe’s own model trails the partners, and the partner models eat credits faster. No specs published for Adobe’s own model. Source ↗

If you already live in Premiere, it’s the shortest path.

Go to Firefly →

Direct link · no commission · * Firefly and Creative Cloud plans include generation credits; partner models cost more credits. We couldn’t load Adobe’s current plan page. · Adobe’s Firefly update ↗

10● NEW

Higgsfield

Length
Depends on the model
Picture
Depends on the model
Sound
Depends on the model
Latest
Sep 21, 2026

Apps & model shops · The ad kit

Higgsfield

From ~$19/mo*Higgsfield AI

Live on the web and app; hosts Veo 3.1, Kling 3.0, Seedance and moreSOME DETAILS UNCONFIRMED

Higgsfield doesn’t make its own video model. It sells the templates around other people’s: ad formats, UGC-style clips, camera moves, and Soul ID for keeping one character’s look across videos. On September 21 it announced new video-ad tools for small businesses, built with OpenAI’s GPT-6 Astra.

What’s new
September 21: small-business video-ad tools, per an OpenAI customer story.
Best for
Ads and UGC-style clips with a consistent character
Where
higgsfield.ai on the web and app, and an MCP connector for AI assistants.

The catch. No model of its own, and the credit math across models is hard to follow. Prices unconfirmed. Source ↗

Good for ads if you like templates. Check the credit math first.

Go to Higgsfield →

Direct link · no commission · * Starter billed annually, 270 credits, per a competitor’s published comparison. Higgsfield’s own pricing page wouldn’t load for us. · OpenAI customer story ↗

11

Pika

Length
Depends on the model
Picture
Up to 1080p
Sound
Yes (Pika Soundtrack)
Latest
Date unconfirmed

Apps & model shops · Cheapest way to try the top models

Pika

From $8/mo*Pika Labs

Live on the web and a social iOS app; hosts Seedance 2.5, MiniMax H3, Wan 3.0 and Omni FlashSOME DETAILS UNCONFIRMED

Pika has quietly become a store for other labs’ models: Seedance 2.5, MiniMax H3, Wan 3.0 and Omni Flash all run inside it, next to its own Pika 2.5. Its new Soundtrack tool adds synced sound effects, voice and music to any clip. At eight dollars a month, it’s the cheapest way we found to try several of the chart leaders side by side.

What’s new
Relaunched as one studio for video, image and sound, plus Pika Soundtrack.
Best for
Trying several top models under one cheap plan
Where
pika.art on the web, the Pika iOS app, and an MCP server.

The catch. Pika’s own model trails the ones it resells. The Pika 2.5 release date and specs come only from aggregators. Source ↗

The tasting menu. Start here if you can’t decide.

Go to Pika →

Direct link · no commission · * Starter billed annually ($10 monthly), 900 credits. · Pika pricing ↗

12

Midjourney Video (V1)

Length
Up to 21 s
Picture
480p (720p HD on Standard and up)
Sound
No
Latest
Jun 18, 2025

Apps & model shops · Moving paintings

Midjourney Video (V1)

From $10/mo*Midjourney

Live on midjourney.com; image-to-video only

Midjourney’s video model is still V1, from June 2025: it takes one of your Midjourney images and sets it moving for a few seconds at a time, up to 21 seconds. No text-to-video and no sound. It stays on the list because nothing else animates the Midjourney look quite like Midjourney does.

What’s new
Nothing new: Midjourney’s updates through September 24 were all image and editor changes. (Some blogs say video shipped in April 2026. It didn’t.)
Best for
Bringing your Midjourney stills to life
Where
midjourney.com on the web. Animates your Midjourney images.

The catch. No audio, no text-to-video, low resolution, and more than a year old. Source ↗

For your own Midjourney art only. Everything else has moved past it.

Go to Midjourney →

Direct link · no commission · * Basic plan, as listed by aggregators; Midjourney’s docs wouldn’t load. HD video needs Standard or higher. · Midjourney V1 video post ↗

13

Grok Imagine 1.5

Length
10 s
Picture
720p
Sound
Unconfirmed
Latest
Jun 3, 2026

Apps & model shops · Fast, inside X

Grok Imagine 1.5

Price not checked*xAI

Live in the Grok apps and the xAI APISOME DETAILS UNCONFIRMED

xAI’s video model is quick: resellers clock about 25 seconds to render a six-second 720p clip. It works best from a still image and lives where Grok users already are.

What’s new
Video 1.5 reached the API in preview June 3, and reportedly became generally available June 16.
Best for
Fast image-to-video in the Grok app
Where
grok.com/imagine, the Grok iOS and Android apps, and the xAI API.

The catch. 720p and 10 seconds. Only the preview date and specs come from xAI; the rest is from aggregators. Source ↗

Fine if you’re already in Grok. Not a reason to start.

Go to Grok Imagine →

Direct link · no commission · * Included in Grok plans; the API price wasn’t confirmed. · xAI’s announcement ↗

14● NEW

LTX-2.5

Length
Not confirmed
Picture
Up to 4K (LTX-2.3 family spec)
Sound
Yes
Latest
Aug 11, 2026

Run it yourself · Fastest on your own GPU

LTX-2.5

Free under $10M revenue*LTX (spun out of Lightricks)

Live as open weights and a managed APISOME DETAILS UNCONFIRMED

LTX is the quick one: about seven seconds to generate a clip on good hardware, with sound, and weights you can download. Version 2.5 adds a new decoder that keeps faces, text and fast motion sharp, and holds a character’s identity across shots.

What’s new
August 11: a sharper decoder and multi-shot identity hold.
Best for
Fast local generation with no per-second meter
Where
Open weights on Hugging Face with ComfyUI support, or the managed LTX API.

The catch. You need a strong GPU. Details come from one news report and aggregators; we found no official 2.5 page. Source ↗

The home-lab pick for speed.

Go to LTX →

Direct link · no commission · * Open weights are free for organizations under $10M in annual revenue; bigger companies need a paid license. Your GPU bill is yours. · LTX blog ↗

15● NEW

MAGI-2 Preview

Length
10 s
Picture
1080p (with the refiner)
Sound
Yes
Latest
Aug 5, 2026

Run it yourself · Truly open, with sound

MAGI-2 Preview

Free (Apache 2.0)*Sand.ai

Live as open weights (Apache 2.0)

MAGI-2 is the highest-ranked fully open model on the with-sound chart, sixth overall. It makes video and sound together, and its Apache 2.0 license has none of the revenue caps attached to H3 or LTX. It’s also a 114-billion-parameter model, so ‘run it yourself’ means real hardware.

What’s new
Open-sourced August 5.
Best for
No-strings open video with sound
Where
Weights on Hugging Face, code on GitHub.

The catch. Ten seconds a clip, and a very large model to host. Source ↗

The open-source benchmark. Bring a big GPU.

Get the weights →

Direct link · no commission · * Open source under Apache 2.0. You pay only for compute. · MAGI-2 on GitHub ↗

THE SPEC SHEET

All fifteen, side by side.

Seen on maker pages and the Artificial Analysis arena · September 29, 2026
ServiceEntry priceLengthPictureSoundStatus
Kling 4.0Just droppedFrom ~$10/mo*30 s (Flash: 20 s)Up to 4K promised (Flash: 720p)Yes, stereoEarly access: 4.0 Flash for Ultra yearly members only. Full model promised for October.
Gemini Omni FlashJust dropped~$0.10/s API*3–10 s per clip (extends to ~40 s)Up to 4KYesLive in the Gemini app, Flow, YouTube Shorts and the Gemini API
Wan 3.0Just dropped$0.05–0.20/s API*30 s1080pYesLive on Alibaba Cloud and the Qwen Cloud API, by application
Seedance 2.5Just droppedFrom $15/mo*30 s (3-minute mode in beta)Up to 4KYesLive in Dreamina (CapCut) worldwide; hosted on Pika, Higgsfield and others
MiniMax H3 (Hailuo 3)Just droppedPrice not published*15 s2KYes, stereo with lip-syncLive in the Hailuo app, on the MiniMax API, and as open weights
Runway Gen-4.5Studio workbenchesFrom $12/mo*Up to a minute, multi-shot1080pYesLive on the web, the app and the API
Luma Ray3.2Studio workbenchesFrom $25/mo*20 s1080p, 16-bit HDR / EXRNo (sound comes from ElevenLabs tools inside Luma)Live in Luma’s app and API
Veo 3.1Studio workbenches$0.05–0.40/s API*8 s (extends to 148 s)Up to 4K (Lite: 1080p)YesLive on the Gemini API, Vertex AI, Flow and the Gemini app
Adobe FireflyStudio workbenchesCredits; price not checked*Depends on the modelDepends on the modelDepends on the modelLive in Firefly and across Creative Cloud
HiggsfieldApps & model shopsFrom ~$19/mo*Depends on the modelDepends on the modelDepends on the modelLive on the web and app; hosts Veo 3.1, Kling 3.0, Seedance and more
PikaApps & model shopsFrom $8/mo*Depends on the modelUp to 1080pYes (Pika Soundtrack)Live on the web and a social iOS app; hosts Seedance 2.5, MiniMax H3, Wan 3.0 and Omni Flash
Midjourney Video (V1)Apps & model shopsFrom $10/mo*Up to 21 s480p (720p HD on Standard and up)NoLive on midjourney.com; image-to-video only
Grok Imagine 1.5Apps & model shopsPrice not checked*10 s720pUnconfirmedLive in the Grok apps and the xAI API
LTX-2.5Run it yourselfFree under $10M revenue*Not confirmedUp to 4K (LTX-2.3 family spec)YesLive as open weights and a managed API
MAGI-2 PreviewRun it yourselfFree (Apache 2.0)*10 s1080p (with the refiner)YesLive as open weights (Apache 2.0)

* See the note on that service’s card. Prices are US list prices before tax and change often; per-second prices are API rates. Monthly plans buy credits, and credits buy fewer seconds on the newest models.

HOW TO READ THE PRICE

Count the seconds,
not the plan.

Almost every plan here sells credits, and every model spends them at a different rate. Runway’s $12 plan buys about 52 seconds of its best model. Luma’s $25 plan buys about 100 seconds at 1080p. The honest unit is the finished second, and on the APIs it’s printed plainly: five cents for Veo 3.1 Lite at 720p, about ten cents for Omni Flash or Wan 3.0 at 720p, forty cents for Veo 3.1 Standard.

And plan for waste. Most clips you generate won’t be the one you keep, so a 30-second model that nails the scene in one take can cost less than a cheap one you roll ten times. That’s the real promise in Kling 4 and Seedance 2.5: keyframes and references mean fewer retries. If you’d rather never watch a meter, MiniMax H3, LTX-2.5 and MAGI-2 will run on your own GPU for the price of the electricity. Already paying for ChatGPT, Gemini or Claude? Our AI plans guide shows which tiers include video.

GONE, COMING, OR NOT YET A PICK

Worth knowing,
not worth a pick.

  • OpenAI SoraClosed

    The Sora app closed April 26, 2026, and the Videos API and every sora-2 model shut down September 24. OpenAI named no replacement, and Sora account data is deleted. Don’t start anything new on it. OpenAI help center ↗

  • Kling 4.0 (full model)Due October

    The 30-second, 4K, 10-keyframe version is promised for October 2026. Until then, only the 720p Flash version is open, to Ultra yearly members. Flash vs. full, explained ↗

  • Seedance 2.5 long modeBeta

    Up to three minutes in one generation. Promising, but it’s labeled beta for a reason. Dreamina ↗

  • HappyHorse 1.0Ranked, little known

    An Alibaba-ATH model sitting fourth on the no-audio chart. We couldn’t find a public product page, so it isn’t a pick yet. Artificial Analysis ↗