A close tabletop shot of a barista pouring a thin stream of steamed milk into a clear glass of dark espresso, forming a crisp rosetta pattern. The hand tilts the pitcher, crema swirls naturally, and a small spoon taps the saucer. Warm window light, locked camera.

Food & Lifestyle · default settings, one run, 5 s, 720p, 16:9 · Updated September 14, 2026

Not run yet
MiniMax FastH3 V2
P-Video 2 Pro
3.8videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency3.5
  • Camera & framing5.0
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio3.0

The locked coffee composition is polished, though the precise barista and rosetta details are less clear.

Not run yet
Wan 3.0
Wan 3.0 Prime
4.2videomodels.ai score
  • Prompt adherence4.0
  • Motion & consistency3.5
  • Camera & framing5.0
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio—

Believable pour and crema, slight pitcher awkwardness.

FastH3 V2
3.9videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency4.0
  • Camera & framing4.5
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio—

The milk pour is clean, but the spoon never taps the saucer.

Seedance 2.5
3.9videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency4.0
  • Camera & framing4.5
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio—

Strong pour and finish, but no spoon tap is visible.

Seedance 2.0
3.8videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency4.0
  • Camera & framing4.5
  • Quality & detail4.0
  • Look & lighting4.5
  • Realism & physics3.5
  • Text rendering—
  • Audio—

Clean locked tabletop pour, but it makes rings rather than a crisp rosetta and no spoon tap is visible.

Seedance 2.0 Fast
3.6videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency3.5
  • Camera & framing4.5
  • Quality & detail4.0
  • Look & lighting4.0
  • Realism & physics3.5
  • Text rendering—
  • Audio—

Milk pours cleanly, but the rosetta remains a faint swirl and the spoon never taps.

Kling 3.0 Turbo
3.5videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency3.0
  • Camera & framing4.0
  • Quality & detail4.0
  • Look & lighting4.0
  • Realism & physics3.5
  • Text rendering—
  • Audio4.0

The warm pour reads well, but there is no clear spoon tap or crisp rosetta.

Veo 3.1
4.0videomodels.ai score
  • Prompt adherence3.5
  • Motion & consistency4.0
  • Camera & framing4.0
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio4.0

Polished rosetta pour, but the spoon tap is not clearly shown.

Veo 3.1 Fast
3.7videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency3.5
  • Camera & framing4.0
  • Quality & detail4.5
  • Look & lighting4.0
  • Realism & physics4.0
  • Text rendering—
  • Audio4.0

It has the closest rosetta-like pour, though the spoon tap never becomes clear.

Happy Horse 1.1
3.9videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency4.0
  • Camera & framing4.5
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio4.0

A strong continuous pour, but the spoon remains static.

MiniMax H3
4.2videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency4.5
  • Camera & framing5.0
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.5
  • Text rendering—
  • Audio4.0

Exceptionally stable and realistic pour, but the spoon never taps.

Gemini Omni Flash
4.2videomodels.ai score
  • Prompt adherence3.5
  • Motion & consistency4.5
  • Camera & framing4.5
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio4.0

Excellent rosetta progression, though the spoon tap is only implied.

Wan 2.7
3.6videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency3.5
  • Camera & framing3.5
  • Quality & detail4.0
  • Look & lighting4.0
  • Realism & physics3.5
  • Text rendering—
  • Audio4.0

Strong cafe image, but the latte art is a spiral rather than a crisp rosetta.

Grok Imagine Video
3.7videomodels.ai score
  • Prompt adherence3.5
  • Motion & consistency4.0
  • Camera & framing4.5
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio1.0

Stable warm scene, but the latte art is not a crisp rosetta.

P-Video 2
4.0videomodels.ai score
  • Prompt adherence3.5
  • Motion & consistency4.0
  • Camera & framing4.0
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio4.0

Warm and polished, but the pattern is a soft spiral rather than a crisp rosetta.

Runway Gen-4.5
3.7videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency3.5
  • Camera & framing4.0
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics3.5
  • Text rendering—
  • Audio—

Attractive pour, but the final hand and pitcher transition jumps.

Hailuo 2.3
3.3videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency3.0
  • Camera & framing4.0
  • Quality & detail3.5
  • Look & lighting3.5
  • Realism & physics3.0
  • Text rendering—
  • Audio—

The pour is clear, but the rosetta is soft and the spoon is moved rather than tapped.

Kling 2.6
3.3videomodels.ai score
  • Prompt adherence2.5
  • Motion & consistency3.0
  • Camera & framing4.0
  • Quality & detail4.0
  • Look & lighting3.5
  • Realism & physics3.5
  • Text rendering—
  • Audio—

The milk pour is clean, but it never produces the requested rosetta or spoon tap.

FLUX 3
3.4videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency2.5
  • Camera & framing3.5
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics3.0
  • Text rendering—
  • Audio4.0

Beautiful latte art and lighting, but the sequence resets after the design forms.

PixVerse V5
3.6videomodels.ai score
  • Prompt adherence2.5
  • Motion & consistency3.5
  • Camera & framing4.0
  • Quality & detail4.5
  • Look & lighting4.0
  • Realism & physics4.0
  • Text rendering—
  • Audio—

The natural milk mixing is convincing, but the requested rosetta and spoon tap are absent.

Seedance 1.5 Pro
3.6videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency3.5
  • Camera & framing4.0
  • Quality & detail4.0
  • Look & lighting4.0
  • Realism & physics3.5
  • Text rendering—
  • Audio4.0

The smooth pour and locked frame work, but its art is a mechanical spiral, not a rosetta.

FastH3
3.8videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency4.0
  • Camera & framing4.5
  • Quality & detail4.0
  • Look & lighting4.5
  • Realism & physics3.5
  • Text rendering—
  • Audio4.0

Clean pour and warm look, but no visible spoon tap or crisp rosetta.

Cosmos 3 Super
3.6videomodels.ai score
  • Prompt adherence2.5
  • Motion & consistency4.0
  • Camera & framing4.5
  • Quality & detail4.0
  • Look & lighting4.0
  • Realism & physics3.5
  • Text rendering—
  • Audio—

Solid locked milk pour, but no crisp rosetta and the spoon tap is unconfirmed.

Lance
3.5videomodels.ai score
  • Prompt adherence1.5
  • Motion & consistency4.0
  • Camera & framing4.5
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics3.5
  • Text rendering—
  • Audio—

Beautifully rendered coffee mixing, but it never makes latte art or the requested espresso setup.

Wan T2V
3.9videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency4.0
  • Camera & framing4.5
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio—

Strong warm locked pour, though the latte art is not a crisp rosetta and tap is unclear.

HunyuanVideo 1.5
3.7videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency3.5
  • Camera & framing4.5
  • Quality & detail4.0
  • Look & lighting4.5
  • Realism & physics3.0
  • Text rendering—
  • Audio—

The pour and locked framing are strong, but it forms rings rather than a rosetta and never shows the spoon tap.

Happy Horse 1.0
4.1videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency4.5
  • Camera & framing4.5
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio4.0

Clean pour and rosetta, but the spoon never taps the saucer.

Helios 14B
3.2videomodels.ai score
  • Prompt adherence2.5
  • Motion & consistency2.5
  • Camera & framing4.0
  • Quality & detail4.0
  • Look & lighting4.0
  • Realism & physics3.0
  • Text rendering—
  • Audio—

The basic pour is present, but the stream flickers, the foam is not a rosetta, and the requested spoon tap never occurs.

Seedance 1.0 Pro Fast
3.0videomodels.ai score
  • Prompt adherence2.5
  • Motion & consistency2.5
  • Camera & framing3.5
  • Quality & detail4.0
  • Look & lighting3.5
  • Realism & physics2.5
  • Text rendering—
  • Audio—

The crema swirl is vivid, but the milk becomes an implausible blob instead of a rosetta.

Wan 2.6
4.1videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency4.5
  • Camera & framing4.5
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio—

Excellent stable pour and warm finish, but the rosetta is generic and no spoon tap occurs.

Kling 3.0 Omni
4.2videomodels.ai score
  • Prompt adherence4.0
  • Motion & consistency4.0
  • Camera & framing4.5
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio—

Best completion of the action chain, with a legible pattern and spoon interaction despite minor artifacts.

P-Video
3.7videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency3.0
  • Camera & framing4.5
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics3.5
  • Text rendering—
  • Audio—

The first half sells the pour well, but it makes concentric rings and then introduces a smeared stirring spoon instead of a saucer tap.

Kling 2.5 Turbo
3.9videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency4.0
  • Camera & framing4.5
  • Quality & detail4.5
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio—

The pour works, but the spoon never taps the saucer.

Seedance 1.0 Pro
3.7videomodels.ai score
  • Prompt adherence2.5
  • Motion & consistency4.0
  • Camera & framing4.5
  • Quality & detail4.0
  • Look & lighting4.5
  • Realism & physics4.0
  • Text rendering—
  • Audio—

The spoon never taps and the latte art is a swirl rather than a rosetta.

PixVerse V4.5
3.8videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency4.0
  • Camera & framing4.5
  • Quality & detail4.0
  • Look & lighting4.0
  • Realism & physics3.5
  • Text rendering—
  • Audio—

Stable warm pour, but the stream is thick and the spoon never taps.

PixVerse V4
3.5videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency3.0
  • Camera & framing4.5
  • Quality & detail4.0
  • Look & lighting4.5
  • Realism & physics3.0
  • Text rendering—
  • Audio—

The coffee pour is clear and attractive, but the oversized drifting steam and nearly static foam reduce realism.

Hailuo 02
3.6videomodels.ai score
  • Prompt adherence3.0
  • Motion & consistency3.5
  • Camera & framing4.5
  • Quality & detail4.0
  • Look & lighting4.0
  • Realism & physics3.5
  • Text rendering—
  • Audio—

Clean locked pour, but the rosetta and spoon-tap actions are not clearly achieved.

Hailuo 01 Director
3.0videomodels.ai score
  • Prompt adherence2.0
  • Motion & consistency2.5
  • Camera & framing4.0
  • Quality & detail4.0
  • Look & lighting4.5
  • Realism & physics2.0
  • Text rendering—
  • Audio—

The pour changes from dark liquid to milk after the latte art is already complete.