AI FILMS Studio cuts Happy Horse 1.1 1080p clips to 945 credits

• AI FILMS Studio cut Happy Horse 1.1's five-second 1080p price from 1,400 credits to 945, a one-third drop according to zaesarius's pricing post.
• In the same post, zaesarius's feature rundown said the model keeps the same capabilities, with text-to-video and image-to-video, 3 to 15 second clips, five aspect ratios, and prompts up to 2,500 characters.
• starks_arq's improvement list and starks_arq's Jupiter notes both framed instruction following, character consistency, and physical realism as the upgrades that mattered most for narrative work.
• The short film Jupiter, shown in starks_arq's film post, used Happy Horse 1.1 as a multi-scene test, and starks_arq's workflow note said the better prompt adherence reduced rework.
• A separate AI FILMS Studio showcase, posted in zaesarius's TILL DAWN spotlight, points to the same platform being used for thriller-style short films as well.

You can open the AI FILMS Studio guide, watch Jupiter, and compare it with the TILL DAWN showcase. The useful bit is not just the cheaper 1080p tier. zaesarius's post also slips in practical workflow details, including named camera moves, aspect-ratio matching for image-to-video, and the suggestion to iterate at 720p before rendering finals at 1080p.

The concrete ship here is simple: 1080p got cheaper without an announced capability downgrade. For narrative creators, that changes the economics of multi-shot iteration faster than a vague quality claim does.

According to zaesarius's pricing post, a five-second 1080p clip fell from 1,400 credits to 945. The same post says Happy Horse 1.1 still offers text-to-video and image-to-video, with 3 to 15 second durations and prompts up to 2,500 characters.

The workflow notes inside zaesarius's feature rundown are unusually concrete:

• Text-to-video supports five aspect ratios.
• The model responds better to named cinematic moves like "slow dolly push," pans, cranes, and rack focus.
• Image-to-video preserves composition and subject identity from the uploaded reference image.
• Output aspect ratio automatically matches the input image in image-to-video mode.
• The model is available in the Nodes Graph Editor.
• The recommended loop is 720p for iteration, 1080p for finals.

The best evidence in the pool is not a benchmark chart. It is a two-minute-plus short film that tries to hold together recurring characters, symbolic scenes, and tonal continuity.

Jupiter, a short film made with Happy Horse 1.1

In starks_arq's setup post, the team said they picked the project because it needed character consistency, emotional performance, and coherence across multiple scenes. starks_arq's follow-up says the biggest gain was instruction following, especially in longer prompts that jump between generations, ideas, and symbolic moments.

The clearest creator-side change is that the upgrade list maps directly to story problems, not just pretty-frame problems. starks_arq's improvement list breaks it into four pieces:

• Better motion and physical realism
• Stronger character consistency
• Improved instruction following
• Higher overall visual quality

starks_arq's workflow note adds the production consequence: less time wrestling with generations, more time refining the film itself. That is the kind of claim that matters when a model is being used for connected scenes instead of isolated clips.

The official-feeling details are buried in zaesarius's tutorial post, not in the creator reactions. They read like a working prompt guide for filmmakers.

According to zaesarius's prompting notes, Happy Horse 1.1 follows explicit cinematic direction better when the prompt names the move instead of describing it loosely. The same post says image-to-video works best when the prompt describes motion, because the frame composition is already supplied by the reference image.

That post also says both text-to-video and image-to-video run inside the Nodes Graph Editor, which turns Happy Horse 1.1 into a workflow component rather than a one-off generator. The linked step-by-step guide is where AI FILMS Studio put the full walkthrough.

One more useful data point is that AI FILMS Studio is already framing the platform around finished short films, not just model demos.

TILL DAWN, a thriller short film highlighted by AI FILMS Studio

zaesarius's TILL DAWN spotlight spotlights Eduardo Bateman's thriller short TILL DAWN and ties it back to the same AI FILMS Studio stack. That does not prove Happy Horse 1.1 made the film, but it does show the platform presenting itself as a place for packaged narrative projects, complete with creator pages and referral codes, not only raw generation tests.