Penjana video AI terbaik: 45 klip, dinilai secara buta

• Diuji pada 1 November 2026
• 5 model
• 3 tugasan, 3 percubaan setiap satu
• 45 klip, 0 gagal
• Data mentah disertakan

Main kelima-limanya, senyap

Prompt yang sama untuk kelima-lima. Ini ialah percubaan 1 daripada 3 bagi setiap model, jadi kami tidak memilih yang kelihatan paling baik. Tekan main pada mana-mana klip.

Jawapan ringkas: jika anda mahukan satu model untuk semua, gunakan Kling 4.0, kerana percubaan terburuknya (3.6) mengatasi percubaan terburuk model lain. Jika harga lebih penting, gunakan Wan 3.0, yang setara dengan Kling secara keseluruhan pada $0.60 sekeping klip. Seedance 2.5 berada betul-betul di belakang mereka dan setara pada klip bercakap. MiniMax H3 memimpin pada adegan kamera bergerak. Pendedahan: Hermoso menjual akses kepada kelima-lima, dan Seedance 2.5 ialah pilihan lalai kami. Ia tidak memenangi mana-mana tugasan secara mutlak.

Siapa patut guna yang mana

Satu model untuk semua

Kling 4.0

Hasil yang paling stabil dalam ujian. Percubaan terendahnya mendapat 3.6, dan tiada tugasan yang menjadi kelemahan. Ia sama tinggi dengan Wan 3.0 untuk tempat pertama keseluruhan.

Nilai terbaik

Wan 3.0

Setara dengan Kling secara keseluruhan pada $0.60 untuk 6 saat, butiran paling tajam dalam ujian dan bunyi semula jadi yang terbaik. MiniMax H3 lebih murah bagi setiap saat, tetapi skornya lebih rendah di luar adegan.

Seseorang bercakap di hadapan kamera

Seri: Wan 3.0, Seedance 2.5, Kling 4.0

Ketiga-tiganya menyebut ayat itu perkataan demi perkataan. Wan memberikan letupan tepung yang paling besar dan lengan berselaput tepung seperti yang kami minta, dengan senyuman iklan stok. Seedance kelihatan paling seperti filem tetapi tersilap kira dalam satu daripada tiga percubaan.

Produk dengan tulisan halus

Kling 4.0, Wan 3.0 hampir

Setiap percubaan Kling mengekalkan ketiga-tiga baris label dengan betul, dengan tangan yang realistik dan gerakan kamera mendekat yang nyata. Ia juga paling menepati taklimat. Seedance menghasilkan madu terbaik dalam ujian, dalam satu percubaan daripada tiga.

Kamera bergerak dalam adegan yang sibuk

MiniMax H3, kemudian Wan dan Kling

Pendedahan yang paling tinggi, dengan penunggang masih dalam bingkai pada akhir. Ia juga mempunyai 8 saat berbanding 6 saat bagi yang lain, yang membantu pergerakan perlahan. Wan 3.0 dan Kling 4.0 berada dalam garis seri kami.

Apabila anda memerlukannya dengan cepat

Veo 3.1

Siap dalam kira-kira 66 saat pada median, berbanding 156 hingga 238 saat bagi yang lain. Hasilnya kelihatan hebat dalam gambar pegun, tetapi ia lebih kerap melangkau langkah dalam taklimat berbanding mana-mana model lain, pada harga tertinggi.

Skor

Setiap klip mendapat 1 hingga 5 pada enam kriteria daripada dua penilai buta yang berasingan. Skor tugasan ialah purata tiga percubaannya, jadi percubaan yang buruk turut dikira. Kami menulis peraturan ini sebelum sebarang render: jurang di bawah 0.3 ialah seri, dan oren menandakan skor tertinggi serta apa-apa yang seri dengannya.

Model	Bercakap	Produk	Adegan	Keseluruhan	Percubaan terburuk
Wan 3.0	4.17	3.67	3.72	3.85	3.29
Kling 4.0	3.95	3.86	3.66	3.82	3.62
Seedance 2.5	4.08	3.33	3.31	3.57	3.00
MiniMax H3	3.36	3.07	3.89	3.44	2.96
Veo 3.1	3.36	2.65	3.10	3.04	2.50

Skor keseluruhan ialah purata biasa bagi 45 skor. Jika audio dikeluarkan sepenuhnya, Kling sedikit mengatasi Wan manakala yang lain kekal di kedudukan masing-masing: Kling 4.0 3.88, Wan 3.0 3.85, Seedance 2.5 3.65, MiniMax H3 3.52, Veo 3.1 3.05. Seedance 2.5 berada dalam garis seri secara keseluruhan, dan percubaan terburuknya ialah sebab kami meletakkannya di belakang dua pemimpin.

Tugasan 1: seseorang bercakap dan mengira dengan jarinya

Seorang tukang roti menyebut satu ayat tepat, menunjukkan satu, dua, kemudian tiga jari, dan menepuk roti supaya tepung naik. Ia menguji penyelarasan bibir, perkataan, anatomi tangan dan fizik kecil dalam satu shot yang tetap. Pemenang: seri tiga hala antara Wan 3.0, Seedance 2.5 dan Kling 4.0. Kesemua 15 klip menyebut ayat itu perkataan demi perkataan. Jari membezakan kedudukan: Veo 3.1 tidak pernah mencapai tiga dalam mana-mana percubaan, dan MiniMax H3 membuka tapak tangan penuh dalam dua daripada tiga percubaan.

Percubaan 1Percubaan 2Percubaan 3

Main kelima-limanya, senyap

Skor bercakap

1. Wan 3.04.17
2. Seedance 2.54.08
3. Kling 4.03.95
4. MiniMax H33.36
5. Veo 3.13.36

Purata tiga percubaan. Oren ialah tempat pertama atau dalam lingkungan 0.3 daripadanya.

Tugasan 2: balang madu dengan label bercetak

Kamera mendekat perlahan pada balang berlabel semasa madu jatuh dari senduk. Ia menguji tulisan kecil, sama ada huruf kekal pegun, dan cecair pekat. Label itu ialah produk rekaan kami sendiri. Pemenang: Kling 4.0, dengan Wan 3.0 dalam garis seri. Ejaan hampir selesai: 14 daripada 15 klip mendapat ketiga-tiga baris dengan betul. Madu tidak. Hampir tiada yang melipat dan melingkar seperti yang diminta, dan Veo 3.1 tidak pernah mendekatkan kamera dalam mana-mana percubaan.

Percubaan 1Percubaan 2Percubaan 3

Main kelima-limanya, senyap

Skor produk

1. Kling 4.03.86
2. Wan 3.03.67
3. Seedance 2.53.33
4. MiniMax H33.07
5. Veo 3.12.65

Purata tiga percubaan. Oren ialah tempat pertama atau dalam lingkungan 0.3 daripadanya.

Tugasan 3: mengikuti seorang penunggang basikal melalui pasar yang hujan

Satu shot tanpa henti di belakang seorang kurier dalam hujan: satu pusingan mengelak troli, seekor anjing melintas, satu deringan loceng, kemudian kamera naik di atas payung untuk menunjukkan seluruh jalan. Pemenang: MiniMax H3, dengan Wan 3.0 dan Kling 4.0 dalam garis seri. Tiada siapa yang mendapat setiap detik. Kling dan H3 paling hampir dengan taklimat. Wan paling tajam tetapi menukar kanvas biasa dengan khemah berwarna dan tanda rekaan. Seedance terus memotong penunggang, dan Veo kehilangan penunggang sepenuhnya dalam dua daripada tiga.

Percubaan 1Percubaan 2Percubaan 3

Main kelima-limanya, senyap

Skor adegan

1. MiniMax H33.89
2. Wan 3.03.72
3. Kling 4.03.66
4. Seedance 2.53.31
5. Veo 3.13.10

Purata tiga percubaan. Oren ialah tempat pertama atau dalam lingkungan 0.3 daripadanya.

Mengikut apa yang kami nilai

Purata merentasi 45 klip. Perincian memisahkan model-model ini paling sedikit. Mengikuti taklimat memisahkannya paling banyak, dan jurang itulah sebahagian besar sebab kedudukan ini kelihatan begitu.

Model yang paling mahal datang terakhir

Veo 3.1 berharga empat kali ganda setiap klip berbanding Wan 3.0 dan mendapat skor terendah dalam setiap tugasan, sama tinggi dengan MiniMax H3 pada klip bercakap. Gambar pegunnya kelihatan cantik, dengan tepung yang paling meyakinkan dalam ujian. Masalahnya ialah langkah: ia tidak pernah mengira sehingga tiga, tidak pernah mendekatkan kamera pada madu, dan kehilangan penunggang dalam hujan. Jika shot anda ialah satu idea yang mudah, ia mungkin lebih sesuai untuk anda daripada yang dicadangkan oleh ujian ini. Prompt kami sengaja penuh dengan langkah.

Ejaan menjadi mudah, cecair tidak

Kami menjangkakan label itu ialah bahagian yang sukar. Empat belas daripada 15 klip mencetak ketiga-tiga baris dengan betul, dan hanya satu percubaan Veo menambah satu perkataan yang tersasar. Madu yang melipat dan melingkar semasa jatuh ialah perkara yang hampir semua orang terlepas. Kebanyakan klip memberi kami lajur lurus, atau bentuk kaku yang hampir tidak berubah. Satu percubaan Seedance 2.5 betul, dan itu juga masalahnya: ia hanyalah satu percubaan daripada tiga.

Tentang muzik pada sebahagian klip madu

Kami meminta bunyi dapur dan tiada muzik. Pada 8 daripada 15 klip madu, model memulangkan hampir tiada bunyi, di bawah -40 dB, dan laluan render kami menambah trek muzik senyap di bawahnya: ketiga-tiga percubaan MiniMax H3, dua Seedance 2.5, dua Veo 3.1 dan satu Wan 3.0. Tiada percubaan Kling yang memerlukannya. Muzik itu milik kami, bukan model. Klip tersebut kehilangan mata pada audio kerana senyap, dan ia ditandakan dalam jadual mentah. Klip bercakap dan adegan tidak disentuh.

Cara kami menjalankannya

Kami menulis tiga prompt dan peraturan pemarkahan dahulu, menjalankan satu klip perintis untuk menyemak saluran, kemudian merender yang lain. Klip perintis ialah salah satu daripada 45. Tiada yang gagal, jadi tiada yang diulang atau digugurkan.

Setiap klip meminta 16:9 pada 720p dan 6 saat, tanpa konteks jenama dan tanpa menulis semula prompt. Dua pengecualian: MiniMax H3 dijalankan pada 8 saat kerana tiada pilihan 6 saat, dan ia menghantar 1344x768 dan bukannya 1280x720. Wan 3.0 menghantar pada 30 bingkai sesaat, yang lain pada 24.

Setiap taklimat memberitahu setiap model perkara yang sama, ditulis mengikut cara yang disyorkan oleh pembuat model itu, kerana ujian yang adil tidak membuatkan satu model membaca dialek model lain. Kling tiada panduan untuk 4.0, jadi kami menggunakan tabiat 3.0nya. Veo 3.1 ialah model penuh, bukan versi lajunya.

Kami menghantar benih 1101, 2202 dan 3303 kepada Veo 3.1, Wan 3.0 dan MiniMax H3. Laluan render kami tidak menghantar benih kepada Seedance 2.5 atau Kling 4.0, jadi percubaan tersebut ialah tiga larian bebas. Tiada model mengesahkan benih yang digunakannya.

Pemarkahan adalah buta dan dilakukan oleh dua penilai AI yang berasingan, bukan manusia. Setiap klip dikurangkan kepada 12 bingkai dari mula hingga akhir, spektrogram bunyinya, kekuatan bunyinya dan transkrip, di bawah kod rawak. Tiada penilai melihat model mana yang menghasilkan klip sehingga skor mereka dikemukakan. Kedua-dua penilai bersetuju dengan rapat, dengan korelasi 0.91 merentasi klip dan jurang purata 0.36 mata bagi setiap kriteria. Skor pada halaman ini ialah purata mereka, dan nombor setiap penilai terdapat dalam CSV.

Beberapa semakan tidak memerlukan pertimbangan. Transkrip tempatan mengesahkan ayat tepat pada kesemua 15 klip bercakap. Pengesanan adegan tidak menemui sebarang pemotongan dalam mana-mana klip, walaupun satu percubaan Veo 3.1 dimainkan sebagai lompatan dan kami menandainya. Kedua-dua penilai mengira baris label secara berasingan dan sepadan pada kesemua 15.

Prompt, perkataan demi perkataan

Salin mana-mana prompt ini. Setiap pautan membuka prompt itu pada model tersebut dalam Hermoso.

A baker speaks to camera in a small bakery kitchen and counts three ingredients on her fingers. She is a woman in her late thirties with short curly dark hair, flour on her forearms and a navy linen apron.

Shot 1 (0-6 seconds): Locked-off medium shot at eye level, 35mm lens, shallow depth of field. She stands behind a floured wooden counter with a round crusty loaf in front of her and looks straight into the lens. She raises her right hand and shows one finger, then two fingers, then three fingers, one finger for each ingredient she names. Then she lowers both hands and pats the top of the loaf twice, and a small puff of flour rises into the window light.

Dialogue: The baker says in a warm, dry voice: "Just three things. Flour. Water. Salt. That's the whole loaf." Accurate lip sync.

Sound: her voice and the soft room tone of the kitchen.

Style: realistic documentary, warm morning window light, one continuous shot.

Negative prompt: subtitles, captions, on-screen text, watermark, background music, cuts.

Medium shot at eye level, locked-off camera, one continuous shot, realistic documentary look, warm morning window light in a small bakery kitchen. A woman in her late thirties with short curly dark hair, flour on her forearms and a navy linen apron stands behind a floured wooden counter with a round crusty loaf in front of her and looks straight into the lens. As she speaks she raises her right hand and shows one finger, then two fingers, then three fingers, one finger for each ingredient she names. Then she lowers both hands and pats the top of the loaf twice, and a small puff of flour rises into the window light.

Baker (warm, dry voice): "Just three things. Flour. Water. Salt. That's the whole loaf."

Only her voice and the soft room tone of the kitchen are heard.

A baker speaks to camera in a small bakery kitchen and counts three ingredients on her fingers, realistic documentary style, locked-off medium shot at eye level, warm morning window light.

0-1 seconds: She stands behind a floured wooden counter with a round crusty loaf in front of her and looks straight into the lens.
1-4 seconds: She raises her right hand and shows one finger, then two fingers, then three fingers, one finger for each ingredient she names.
4-6 seconds: She lowers both hands and pats the top of the loaf twice, and a small puff of flour rises into the window light.

Dialogue: The baker says in a warm, dry voice: "Just three things. Flour. Water. Salt. That's the whole loaf."

Constant throughout: she is a woman in her late thirties with short curly dark hair, flour on her forearms and a navy linen apron. One continuous shot, the camera stays still. Sound: her voice and the soft room tone of the kitchen. No subtitles. No background music.

integrated_multimodal_description: [Shot 1] Live-action, realistic documentary style, a Static Shot at eye level holds a medium shot of a small bakery kitchen in warm morning window light. A woman in her late thirties with short curly dark hair, flour on her forearms and a navy linen apron (S1), on screen, with a warm, dry voice, stands behind a floured wooden counter with a round crusty loaf in front of her. She looks straight into the lens and says: <d>[English] Just three things. Flour. Water. Salt. That's the whole loaf.</d> As she speaks she raises her right hand and shows one finger, then two fingers, then three fingers, one finger for each ingredient she names. Then she lowers both hands and pats the top of the loaf twice, and a small puff of flour rises into the window light.

overall_soundscape: The soft room tone of a small kitchen. Her palms pat lightly on the bread crust.

non_diegetic_music: N/A

A medium shot at eye level with a locked-off camera in one continuous take. In a small bakery kitchen lit by warm morning window light, a woman in her late thirties with short curly dark hair, flour on her forearms and a navy linen apron stands behind a floured wooden counter with a round crusty loaf in front of her and looks straight into the lens. She raises her right hand and shows one finger, then two fingers, then three fingers, one finger for each ingredient she names. She then lowers both hands and pats the top of the loaf twice, and a small puff of flour rises into the window light. Realistic documentary style, 35mm lens, shallow depth of field.

The baker says in a warm, dry voice: "Just three things. Flour. Water. Salt. That's the whole loaf."

Audio: her voice and the soft room tone of the kitchen, and nothing else.

A macro product shot of an open jar of honey with a printed label while a ribbon of honey falls into it.

Shot 1 (0-6 seconds): Slow push-in, 100mm macro lens. The frame shows an open, squat, clear glass jar of golden honey on a dark slate counter in soft daylight from the left. A cream paper label on the front of the jar is printed in black with exactly three lines of text: "HERMOSO" in large serif capitals, "WILDFLOWER HONEY" in smaller capitals below it, and "NET WT 12 OZ" in small capitals at the bottom. A hand at the top of the frame holds a wooden honey dipper above the jar, and a thin ribbon of thick honey falls into the open jar, where it folds and coils on the surface before slowly sinking. The camera pushes in slowly toward the label until the three lines of text fill the lower third of the frame, sharp and legible, with the letters fixed in place on the glass.

Sound: a quiet kitchen with a kettle simmering softly and birdsong through an open window.

Style: photoreal commercial, soft daylight, one continuous shot.

Negative prompt: subtitles, captions, watermark, background music, extra text on the label, cuts.

Macro product shot, one continuous shot, photoreal commercial look. The frame shows an open, squat, clear glass jar of golden honey on a dark slate counter in soft daylight from the left. A cream paper label on the front of the jar is printed in black with exactly three lines of text: "HERMOSO" in large serif capitals, "WILDFLOWER HONEY" in smaller capitals below it, and "NET WT 12 OZ" in small capitals at the bottom. A hand at the top of the frame holds a wooden honey dipper above the jar, and a thin ribbon of thick honey falls into the open jar, where it folds and coils on the surface before slowly sinking. The camera pushes in slowly toward the label until the three lines of text fill the lower third of the frame, sharp and legible, with the letters fixed in place on the glass.

Sound: a quiet kitchen with a kettle simmering softly and birdsong through an open window.

A macro product shot of an open jar of honey with a printed label while a ribbon of honey falls into it, photoreal commercial style, slow push-in, soft daylight.

0-2 seconds: The frame shows an open, squat, clear glass jar of golden honey on a dark slate counter in soft daylight from the left. A cream paper label on the front of the jar is printed in black with exactly three lines of text: "HERMOSO" in large serif capitals, "WILDFLOWER HONEY" in smaller capitals below it, and "NET WT 12 OZ" in small capitals at the bottom.
2-4 seconds: A hand at the top of the frame holds a wooden honey dipper above the jar, and a thin ribbon of thick honey falls into the open jar, where it folds and coils on the surface before slowly sinking.
4-6 seconds: The camera keeps pushing in slowly toward the label until the three lines of text fill the lower third of the frame, sharp and legible.

Constant throughout: the label's letters stay fixed in place on the glass. One continuous shot. Sound: a quiet kitchen with a kettle simmering softly and birdsong through an open window. No subtitles. No background music.

integrated_multimodal_description: [Shot 1] Live-action, photoreal commercial style, a macro shot shows an open, squat, clear glass jar of golden honey on a dark slate counter in soft daylight from the left. A cream paper label on the front of the jar is printed in black with exactly three lines of text: "HERMOSO" in large serif capitals, "WILDFLOWER HONEY" in smaller capitals below it, and "NET WT 12 OZ" in small capitals at the bottom. A hand at the top of the frame holds a wooden honey dipper above the jar, and a thin ribbon of thick honey falls into the open jar, where it folds and coils on the surface before slowly sinking. The camera performs a Push In at slow speed toward the label until the three lines of text fill the lower third of the frame, sharp and legible, with the letters fixed in place on the glass.

overall_soundscape: A quiet kitchen with a kettle simmering softly. Birdsong drifts through an open window.

non_diegetic_music: N/A

A macro product shot in one continuous take with a slow push-in, 100mm macro lens. The frame shows an open, squat, clear glass jar of golden honey on a dark slate counter in soft daylight from the left. A cream paper label on the front of the jar is printed in black with exactly three lines of text: "HERMOSO" in large serif capitals, "WILDFLOWER HONEY" in smaller capitals below it, and "NET WT 12 OZ" in small capitals at the bottom. A hand at the top of the frame holds a wooden honey dipper above the jar, and a thin ribbon of thick honey falls into the open jar, where it folds and coils on the surface before slowly sinking. The camera pushes in slowly toward the label until the three lines of text fill the lower third of the frame, sharp and legible, with the letters fixed in place on the glass. Photoreal commercial style.

Audio: a quiet kitchen with a kettle simmering softly and birdsong through an open window, and nothing else.

A low tracking shot follows a bicycle courier through a crowded rainy night market and rises above the crowd to reveal the whole street.

Shot 1 (0-6 seconds): Wide 24mm lens. In a narrow night-market street at dusk in heavy rain, lined with food stalls under plain canvas awnings and strings of warm lanterns, a bicycle courier in a yellow rain jacket rides away from the camera while the camera follows behind at low height at the same speed. The rider swerves left around a vendor pushing a cart, then a small dog runs across the street in front of the bicycle and the rider rings the bell once. Steam rises from a noodle stall and the wet cobblestones reflect the lanterns. In the last two seconds the camera rises smoothly above a crowd of people with umbrellas and reveals the whole street stretching ahead, rain streaks catching the lantern light.

Sound: heavy rain on canvas, one bicycle bell, sizzling woks, crowd chatter.

Style: cinematic realistic, rainy dusk, one continuous shot.

Negative prompt: subtitles, captions, on-screen text, watermark, background music, cuts.

One continuous tracking shot, cinematic realistic look. In a narrow night-market street at dusk in heavy rain, lined with food stalls under plain canvas awnings and strings of warm lanterns, a bicycle courier in a yellow rain jacket rides away from the camera while the camera follows behind at low height at the same speed. The rider swerves left around a vendor pushing a cart, then a small dog runs across the street in front of the bicycle and the rider rings the bell once. Steam rises from a noodle stall and the wet cobblestones reflect the lanterns. In the last two seconds the camera rises smoothly above a crowd of people with umbrellas and reveals the whole street stretching ahead, rain streaks catching the lantern light.

Sound: heavy rain on canvas, one bicycle bell, sizzling woks, crowd chatter.

A low tracking shot follows a bicycle courier through a crowded rainy night market and rises above the crowd to reveal the whole street, cinematic realistic style, one continuous shot.

0-2 seconds: In a narrow night-market street at dusk in heavy rain, lined with food stalls under plain canvas awnings and strings of warm lanterns, a bicycle courier in a yellow rain jacket rides away from the camera. The camera follows behind at low height at the same speed. The rider swerves left around a vendor pushing a cart.
2-4 seconds: Ahead, a small dog runs across the street in front of the bicycle and the rider rings the bell once. Steam rises from a noodle stall and the wet cobblestones reflect the lanterns.
4-6 seconds: The camera rises smoothly above a crowd of people with umbrellas and reveals the whole street stretching ahead, rain streaks catching the lantern light.

Constant throughout: the same rider, the same rainy dusk light. Sound: heavy rain on canvas, one bicycle bell, sizzling woks, crowd chatter. No subtitles. No background music.

integrated_multimodal_description: [Shot 1] Live-action, cinematic realistic style, a Tracking Shot at low height follows a bicycle courier in a yellow rain jacket who rides away from the camera through a narrow night-market street at dusk in heavy rain, lined with food stalls under plain canvas awnings and strings of warm lanterns, with the camera moving at the same speed. The rider swerves left around a vendor pushing a cart, then a small dog runs across the street in front of the bicycle and the rider rings the bell once. Steam rises from a noodle stall and the wet cobblestones reflect the lanterns. Near the end the camera performs a Pedestal Up with large amplitude at slow speed, rising above a crowd of people with umbrellas and revealing the whole street stretching ahead, rain streaks catching the lantern light.

overall_soundscape: Heavy rain drums on canvas and one bicycle bell rings. Woks sizzle under the crowd's chatter.

non_diegetic_music: N/A

One continuous tracking shot with a wide 24mm lens, cinematic realistic style. In a narrow night-market street at dusk in heavy rain, lined with food stalls under plain canvas awnings and strings of warm lanterns, a bicycle courier in a yellow rain jacket rides away from the camera while the camera follows behind at low height at the same speed. The rider swerves left around a vendor pushing a cart, then a small dog runs across the street in front of the bicycle and the rider rings the bell once. Steam rises from a noodle stall and the wet cobblestones reflect the lanterns. In the last two seconds the camera rises smoothly above a crowd of people with umbrellas and reveals the whole street stretching ahead, rain streaks catching the lantern light.

Audio: heavy rain on canvas, one bicycle bell, sizzling woks and crowd chatter, and nothing else.

Apa yang ujian ini tidak dapat beritahu anda

Tiga prompt tidak meliputi segalanya, dan tiga percubaan bagi setiap tugasan ialah sampel yang kecil, jadi anggap jurang 0.3 atau kurang sebagai tiada jurang langsung. Kedua-dua penilai ialah model AI yang membaca bingkai dan ukuran. Tiada seorang pun mendengar klip ini dengan telinga: bunyi dinilai daripada transkrip, spektrogram dan kekuatan bunyi, jadi kualiti suara ialah bahagian paling lemah dalam pemarkahan kami, dan 12 bingkai boleh terlepas gangguan di antara bingkai-bingkai itu. Ukuran penyelarasan bibir kami kasar dan kami tidak bergantung padanya. Ia meletakkan Seedance 2.5 dan Kling 4.0 paling hampir, dan Wan 3.0 sehingga setengah saat tersasar, jadi tontonlah klip jika penyelarasan bibir penting bagi anda. Ujian buta mempunyai satu lubang yang diketahui: panjang 8 saat MiniMax H3 menjadikan klipnya mudah dikenali, dan dua saat tambahan itu mungkin membantunya pada pendedahan yang perlahan. Apa-apa yang tiada dalam jadual, termasuk Seedance 2.0 dan HeyGen Video, tidak diuji. Model dan harga berubah dengan cepat, dan keputusan ini datang daripada satu hari sahaja.