Head to head: Happy Horse 1.1 Image to Video vs LTX 2.3 Video Fast
Happy Horse 1.1 Image to Video vs LTX 2.3 Video Fast
By Ryan Merket · Published
These two video models split along clear practical lines: one favors subject fidelity and stable motion, while the other is stronger at sustained camera movement. The aggregate scores alone do not tell the full story.
Happy Horse posts the higher aggregate score, 29.9 to 26.1, but that margin is not statistically decisive. With only limited confidence that either model is genuinely better, the overall result is an effective tie—not a narrow victory dressed up as certainty. The task breakdown is more useful. Happy Horse repeatedly leads in slow-motion detail and temporal consistency: its hummingbirds remain recognizable around scarlet heliconias, and its juvenile coati holds its anatomy, markings, ringed tail, and rear-to-profile progression more reliably. It also generally handles balloon-burst imagery with clearer narrative staging and polished water-sheet expansion, although some crown-like fluid geometry looks conspicuously artificial; one fluid comparison was a tie, and LTX sometimes produced the more natural droplet dispersion. LTX owns the single-continuous-shot tests. Its cathedral sequences communicate the glide from dark entrance to altar more explicitly and consistently, even when heavy orange grading, blown candle highlights, and softer architecture cost it visual refinement. Happy Horse often renders the prettier frame, but LTX executes the requested camera journey more convincingly. That makes this a workload decision, not a podium finish: choose Happy Horse for stable creatures, prompt-specific detail, and polished slow motion; choose LTX for legible forward camera movement and continuous-shot progression. **Final call: too close to call—an overall tie, with only limited confidence that either model is truly better.**
Fluid & particle dynamics
Cinematic slow-motion of a water balloon bursting, the sheet of water expanding and droplets scattering through the air in convincing detail against a dark background, hard side light, 16:9.
OpenAI: GPT-5.6 Sol Pro: Model A clearly depicts a balloon bursting into an illuminated water sheet with strong cinematic polish, though the collapsing red material and crown-like water geometry look somewhat artificial. Model B offers more convincing droplet dispersion and temporal progression, but the balloon itself is absent, weakening prompt adherence and narrative clarity. (Second judge pass, order swapped — scores are the average of both: Model B lacks a clearly visible balloon but presents a more convincing, temp | Anthropic: Claude Fable 5.1: Model A shows the red balloon with nice ha…
Single continuous shot
One unbroken take gliding slowly through a candlelit cathedral from the entrance toward the altar, no cuts, jumps, or transitions, dust and warm light in the air, 16:9.
OpenAI: GPT-5.6 Sol Pro: Model A presents a highly coherent slow forward glide with realistic cathedral detail, warm candlelight, visible atmospheric haze, and excellent spatial consistency. Model B clearly progresses from the entrance toward the altar, but its heavy orange cast, blown-out candles, and softer architectural detail reduce visual fidelity and refinement. (Second judge pass, order swapped — scores are the average of both: Model B shows clearer forward progression toward the altar and stronger candlelit warm | Anthropic: Claude Fable 5.1: Model B clearly executes the gliding dolly…
Slow-motion detail
A single continuous 6-second shot in 16:9 of a violet-crowned woodnymph hummingbird braking in midair to sip from a scarlet heliconia after a sudden tropical raindrop strikes the flower cup, sending a crystalline crown of nectar and water outward in convincing slow motion; the camera executes a gentle push-in with a slight upward tilt from close-up to tighter macro as the bird hovers, tail fans, and wingbeats remain crisp and fluid frame to frame, with razor-detailed droplets, feather microtexture, and rippling flower petals; luminous overcast jungle light with a bright rim from a clearing behind, hushed awe-filled mood.
OpenAI: GPT-5.6 Sol Pro: Model A better captures the specified woodnymph hovering beside a recognizable scarlet heliconia, with convincing wing positions, jungle rim light, and polished detail, though the splash is less crown-like and the flower shifts somewhat between frames. Model B presents a stronger macro droplet burst and good continuity, but the flower is not clearly a heliconia, while the cropped bird, obscured wings, and absent tail fan weaken prompt adherence and motion readability. (Second judge pass, order s | Anthropic: Claude Fable 5.1: Model A delivers a convincingly realistic…
Temporal consistency
A single continuous 8-second shot in 16:9 of one juvenile ring-tailed coati with a distinct broken white eyebrow stripe, a nick in its left ear, and a faded teal tracking collar trotting along a mossy fallen ceiba trunk above a cloud-forest ravine; the camera begins low behind the animal and performs a smooth side-dolly that arcs slightly forward into a medium profile as it climbs, pauses to sniff bromeliads, and looks straight toward lens, while its fur pattern, collar fit, facial markings, body size, and identity remain perfectly unchanged with no morphing or flicker; cool dawn mist, shafts of pale sunlight through leaves, tiny gnats drifting, alert exploratory mood.
OpenAI: GPT-5.6 Sol Pro: Model A more clearly delivers the requested rear-to-profile-to-lens progression, bromeliad sniffing, ringed tail, misty lighting, and stable juvenile identity. Model B is attractive and coherent, but its facial markings are less distinctive and the tail pattern and body appearance vary more noticeably across frames. (Second judge pass, order swapped — scores are the average of both: Model A more clearly depicts a juvenile ring-tailed coati and better matches the fallen-trunk traversal, bromeliad | Anthropic: Claude Fable 5.1: Model A renders a convincing coati with su…
Matchup powered by OpenRouter.