Skip to content
ReviewRamble

The weekly brief

AI, reviewed, in your inbox.

  • New reviews of tools, models, MCP servers and repos, with the verdict up front
  • Re-checked prices and ratings when something changes
  • New analysis pieces as they are published

Models, agents, MCP servers and the tools actually worth your time.

One email a week. Unsubscribe with the link in any brief. How we use your email

HiDream-O1-Video-1.0

HiDream's first video model promised text, image and video in with 20-second 1080p out; the public API takes one image.

Reviewed by ReviewRamble editors · Updated

Share this page:XLinkedInFacebookRedditWhatsAppEmail

Our rating

3.4/5

HiDream-O1-Video-1.0 launched on a press release's promises and ships as a single-image animator that has slipped down the boards.

What is HiDream-O1-Video-1.0?

HiDream-O1-Video-1.0 is the first video model from HiDream.ai (ZhiXiang Future), released in China on September 15, 2026, with an international announcement on September 17. HiDream calls it a native omnimodal model that represents text, video and audio in one framework, and says it generates 5 to 20 second 1080p clips with synchronised sound from text, image or video input, choosing the clip length to suit the action and modelling gravity, inertia and collisions. It belongs to a family of models on HiDream's Unified Transformer, alongside O1-Image, O1-World and O1-Embodied. No weights or parameter count have been published.

The public route is narrower. HiDream's HiHarness API documents it as single-image-to-video: one reference image of 40 KB to 20 MB plus an optional prompt, with output matching the image's aspect ratio and a length either chosen by the model or fixed at 10 seconds. Creators can use it in vivago R1 Studio, where it went live on September 30.

The same HiHarness docs already list a HiDream-O1-Video-1.1 model ID with text-to-video, first-and-last-frame and reference modes, 5 to 16 seconds and 480p to 4K output, which HiDream had not announced as of November 21.

Is HiDream-O1-Video-1.0 worth it?

HiDream-O1-Video-1.0 is a smaller product than its announcement. HiDream described text, image and video input and 5 to 20 second clips at 1080p with sound; the API developers can call takes one still image and a prompt, offers no resolution setting and no audio control, and lets you choose only between the model's own length and exactly 10 seconds.

Its standing has faded too. At launch HiDream claimed 4th on Artificial Analysis's image-to-video board, and in October it sat 6th at 1175 Elo on the board's earlier version. On the current version it is 14th at 1103, behind Pika H3 Flex and Agnes Video 2.5, and it does not appear on the text-to-video board at all.

The catch: even the basics need care. A finished HiHarness task reports status 1 when every subtask has ended, not when one has succeeded, so a client has to read each subtask's own status to tell a video from a safety rejection. HiDream publishes no rate card of its own; Artificial Analysis's $5.80 a minute is the only price on record.

Our take: the physics-aware claims may be real, and the undocumented 1.1 in the API suggests the full model is on its way. Until it arrives, Pika H3 Flex animates a still for less and ranks higher.

Where it earns its place

  • A concept artist at a small games studio

    See a painted environment come alive for a pitch deck

    Upload the painting to vivago R1 Studio or HiHarness with a short motion prompt and let the model pick the clip length; set force_10s for a fixed 10-second loop.

  • A developer evaluating physics-heavy image-to-video

    Compare falling, colliding and pouring motion across vendors

    Run the same stills through HiHarness and a rival, checking each subtask's status rather than the task's, since status 1 only means the job ended.

Before you commit

Setup
Get a HiHarness API key, POST one image (URL or Base64, 40 KB to 20 MB) with the HiDream-O1-Video-1.0 model ID, then poll the results endpoint with the task_id.
Integrations
HiHarness asynchronous REST API; vivago R1 Studio for creators; OrcaRouter does not serve it.
Data and privacy
HiDream's retention and training terms for uploaded images were not published on the API documentation we read.
Cost at scale
HiDream publishes no rate card; at Artificial Analysis's $5.80 a minute, 100 minutes a month would be $580.
Limits
One input image; output aspect ratio follows the image; length model-chosen or fixed at 10 seconds; check sub_task_results for failures and safety rejections.

What’s good

  • Image-to-video with sound at about $5.80 a minute on Artificial Analysis's figures
  • Model chooses clip length to fit the action
  • Part of a wider HiDream family with image and world models

What’s not

  • Public API takes one image only, with no resolution or audio controls
  • 14th on Artificial Analysis's current image-to-video board, down from 6th in October
  • No official rate card, weights or parameter count

Best for

Teams in HiDream's ecosystem who want to animate single stills and are willing to wait for the fuller 1.1 release.

HiDream-O1-Video-1.0 alternatives

HiDream-O1-Video-1.0: common questions

Is HiDream-O1-Video-1.0 worth it?
HiDream-O1-Video-1.0 launched on a press release's promises and ships as a single-image animator that has slipped down the boards. We rate it 3.4 out of 5.
How much does HiDream-O1-Video-1.0 cost?
HiHarness API; HiDream publishes no per-clip rate card. Artificial Analysis lists $5.80 per minute.
What are the downsides of HiDream-O1-Video-1.0?
Public API takes one image only, with no resolution or audio controls; 14th on Artificial Analysis's current image-to-video board, down from 6th in October; no official rate card, weights or parameter count.
Who is HiDream-O1-Video-1.0 best for?
Teams in HiDream's ecosystem who want to animate single stills and are willing to wait for the fuller 1.1 release.
How much does HiDream-O1-Video-1.0 cost for a team or in production?
HiDream publishes no rate card; at Artificial Analysis's $5.80 a minute, 100 minutes a month would be $580.
Is HiDream-O1-Video-1.0 safe for company data?
HiDream's retention and training terms for uploaded images were not published on the API documentation we read.
What are the best HiDream-O1-Video-1.0 alternatives?
The closest alternatives we have reviewed are Pika H3 Flex, MiniMax H3 (Hailuo 3.0), Vidu Q4 Preview, Wan 3.0 and Agnes Video 2.5.

Answers checked Nov 21, 2026.