Genativity

Benchmark

Hands leaderboard

Every model runs the same 4 prompts on this board. A vision model scores each clip blind against the original prompt in a single pass, across visual fidelity, physics and motion, subject consistency, prompt adherence and audio sync. Scores are out of 5, and the 95% CI column shows how much a score could plausibly move on a re-run with this few samples. Full methodology.

Board: Overall · Hands · Faces · Liquids · On-screen text · Multi-shot · Lip-sync · Camera · Animals · Style · Image to video · Animation

Look: Realism · Animation and 3D

Round: all · v1-vivix-2026-10 · v1-2026-10 · v1-2026-09 · anim1-2026-10

Clip shown: every model here on the same prompt, “hands-tying-shoelaces”, so the preview is a fair comparison, not a cherry-picked best clip.

#ModelClipOverall95% CIFidelity VisualPhysicsConsistency AudioAttemptsGen (s) $/runRuns
1 MiniMax H3
MiniMax
4.83 ±0.53 4.63 5.00 4.88 — — — 448 — 4
2 OpenAI Sora 2 discontinued
OpenAI
4.75 ±0.46 4.63 4.63 5.00 — — 1.00 217 — 4
3 Seedance 2.5
ByteDance
4.67 ±0.75 4.50 4.75 4.75 — — — 233 — 4
4 Google Veo 3.1
Google
4.58 ±1.33 4.38 4.75 4.63 — — — 132 — 4
5 Kling 4.0 preview
Kuaishou
4.58 ±0.8 4.13 4.88 4.75 — — — — — 4
6 Kling 3.0 Turbo
Kuaishou
4.53 ±1.41 4.25 4.50 4.38 4.50 5.00 — — — 4
7 FLUX 3
Black Forest Labs
4.46 ±0.85 4.50 4.75 4.13 — — 1.00 102 $1.360 4
8 Wan 3.0 preview
Alibaba
4.37 ±0.85 4.25 4.75 4.13 — — — — — 4
9 PixVerse v5.5
PixVerse
4.22 ±1.69 3.88 4.50 4.13 4.38 — — — — 4
10 Seedance 2.0
ByteDance
4.09 ±1.46 3.75 4.38 4.13 — — — 212 — 4
11 Google Gemini Omni
Google
4.01 ±1.32 3.83 4.20 3.63 3.58 4.80 — — — 4
12 Kling 3.0
Kuaishou
3.96 ±1.43 3.88 4.38 3.63 — — 1.00 190 — 4
13 Grok Video
xAI
3.88 ±1.51 3.75 4.25 3.63 — — 1.00 85 — 4
14 MiniMax Hailuo 02
MiniMax
3.88 ±0.7 4.00 3.88 3.75 — — — 300 — 4
15 Vivix W1 preview
Vivix
3.83 ±1.51 3.50 4.25 3.75 — — 1.00 18 $0.027 4
16 Runway Gen-4.5
Runway
3.53 ±0.44 3.38 4.00 3.13 3.63 — — — — 4

Attempts is the average number of generations needed before one was usable, the number that actually decides what a model costs you. 95% CI is computed from that model's own runs on this board, not assumed. On a category board with only 4 runs per model, a wide CI is the honest reflection of a small sample, not a bug. Two models whose CIs overlap are not meaningfully different at this sample size.

Cite this

Free to cite and reuse with a link back. Data is updated as new runs and prices come in, so include the date.

Genativity. (2026). Hands leaderboard: AI video models ranked. Retrieved October 23, 2026, from https://genativity.com/leaderboard/hands
<a href="https://genativity.com/leaderboard/hands">Hands leaderboard: AI video models ranked</a> (Genativity)

Frequently asked

Which AI video model is best at Hands right now?

As of 23 October 2026, MiniMax H3 by MiniMax ranks #1 on Hands on Genativity's benchmark, scoring 4.83/5 across 4 scored runs (±0.53 at 95% confidence).