Reka - Most evaluations of world models ask, "Does this look right?" 🌎 We b
Reka · Jul 2026 · 0:59 · AI-annotated
- Length
- 0:59
- Scenes
- 34 · 1.7 s avg
- Product on screen
- 0:30
- Call to action
- 0:54 · 91%
- Voiceover
- No
Reka’s 0:59 launch film runs 34 scenes, about 1.7 s each. The product shows up at 0:30, with social proof. The call to action lands at 0:54 (91% in). Short beats stack on a black void, alternating photoreal 3D renders, teal dot-matrix dissolves, and flat cyan monospace cards, cut on a steady pulse. Signature moves are the halftone dissolve that morphs renders into dot fields and the cyan panel wipe that swaps footage for text.
Is this your launch? to your site or press kit.
More like this
The Manufacturing Intelligence Company · Jul 2026
NeoSigma · Nov 2026
Reka - Most evaluations of world models ask, "Does this look right?" 🌎 We b, scene by scene
· 34 scenes as text
01 · Hook · 0:00–0:02
A red egg sits in a circular recess on a black platform against a black void. A white robotic hand enters from the right, hovers over the egg and curls its fingers, while the background fills with a faint glitchy digital texture.
Camera: Static
02 · Hook · 0:02–0:03
A hand built from glowing teal dots reaches down from the right over a flat surface, fingers curling toward a small sphere resting on it, all set on a black background.
Techniques: Particles
Camera: Pull out
03 · Hook · 0:03–0:05
A teal-and-white dot-matrix shape, curving like a hand or arm, rests in the lower third on a black field. A horizontal line of dots anchors the bottom edge. The dots flicker and shift in place, giving the form a shimmering, dithered motion.
Techniques: Particles
Camera: Push in
04 · Other · 0:05–0:08
A robotic arm rendered in green and cyan halftone dots moves across the lower right of a black field, its dot matrix shifting as it rises, while a white monospace caption appears above it.
On-screen text: “A world model is a learned simulator.”
Techniques: 3D mockup
Camera: Pull out
05 · Other · 0:08–0:09
A thin white rectangular window on a black background frames a short clip of a hand reaching for a small object on a table. A green halftone shape dissolves into the footage as the small monospaced caption reveals across the frame.
On-screen text: “A world model is a learned simulator.”
Techniques: Masking
Camera: Pull out
06 · Other · 0:09–0:16
On black, a diagram assembles left to right: a photo card labeled state_t holds a robotic hand, an action line feeds a cyan MODEL box, and a second card fills with green dots resolving into the hand. A dashed frame draws around the group, then the layout tilts into 3D.
On-screen text: “state_t”, “action”, “[ MODEL ]”, “state_{t+1}”, “Given a state and an action, it predicts what happens next.”
Techniques: Morph, Masking
Camera: Push in
07 · Other · 0:16–0:17
A cyan rectangle scales up at center on a black field while thin monospace text inside deletes from the left, MODEL]] then EL]] then ]. Out-of-focus 3D panels and faint perspective grid lines sit behind it.
On-screen text: “MODEL]]”
Techniques: Kinetic type
Camera: Push in
08 · Other · 0:17–0:17
Two hands holding a tablet emerge from a cyan halftone dot field, the dots densifying into a crisp white pattern, then the word Agents fades in at the bottom.
On-screen text: “Agents”
Camera: Push in
09 · Other · 0:17–0:19
Two white robotic hands with dark joints cup a plain white cube in 3D perspective. The soft background shifts from cool blue to warm orange while a faint dotted globe fades behind. The caption stays fixed at the bottom.
On-screen text: “Agents plan inside them”
Techniques: 3D mockup
Camera: Pull out
10 · Other · 0:19–0:20
Two white robotic hands grip a matte white cube held at the center of a dark frame, a soft diagonal beam of light falling from the upper left. The caption sits still at the bottom while the hands keep their hold on the cube.
On-screen text: “Look ahead with them”
Techniques: 3D mockup
Camera: Static
11 · Other · 0:20–0:21
Two white robotic arms with black joints hold a small white rectangular device at the center of frame against a warm-to-dark gradient background. The hands cradle the object steadily while a dark caption bar sits along the bottom edge.
On-screen text: “Look ahead with them”
Techniques: 3D mockup
Camera: Static
12 · Other · 0:21–0:21
Two white robotic hands hold a small white cube against a hazy gradient. A blue field wipes up across the frame, dimming the white line at the bottom as the orange text “All on one” appears at the left.
On-screen text: “Look ahead with them”, “All on one”
Techniques: Masking
Camera: Static
13 · Problem · 0:21–0:22
A cyan panel wipes up from the bottom to cover a 3D rendered device at the top of the frame, leaving the black headline centered on a plain cyan field.
On-screen text: “All on one assumption”
Techniques: Masking
Camera: Static
14 · Problem · 0:22–0:23
A cyan title card holding a centered monospace line wipes upward to reveal a workshop bench, where two white robot grippers hold a taped cardboard box in front of a tool wall.
On-screen text: “All on one assumption”
Techniques: Masking
Camera: Pan
15 · Problem · 0:23–0:26
From the robot's first-person view, two white robotic hands hold a taped cardboard box and lower it onto a wooden workbench, then release their grip and pull away. The camera tracks the descent. Shelves of bins, tools and wires fill the workshop background.
Camera: Push in
16 · Problem · 0:26–0:28
Black monospace text on a cyan field reads that the imagined world as a second line drops in. A darker cyan panel then wipes over the frame, revealing the line That assumption is rarely tested. in light cyan.
On-screen text: “that the imagined world”, “behaves like the real one.”, “That assumption is rarely tested.”
Techniques: Masking
Camera: Static
17 · Problem · 0:28–0:30
A black field framed by a thin cyan border that drops away within the first half second. A centered white monospace caption holds the line, its words settling from a slight vertical stagger into a clean row, then holding on a locked-off frame.
On-screen text: “That assumption is rarely tested.”
Techniques: Kinetic type
Camera: Static
18 · Product intro · 0:30–0:31
A mosaic of small pixelated environment thumbnails fills the frame behind a centered white title card. The composition starts soft and small, then racks into focus and scales up as the card settles, holding the title and subtitle.
On-screen text: “WorldModelGym”, “test it directly.”
Techniques: Kinetic type
Camera: Pull out
19 · Product intro · 0:31–0:33
A mosaic of retro game and robotics environment thumbnails fills the frame behind a centered white card reading “WorldModelGym / test it directly.” The grid defocuses into a blur as a white “Take” caption fades in at the bottom left.
On-screen text: “WorldModelGym”, “test it directly.”, “Take”
Techniques: Kinetic type
Camera: Static
20 · Demo · 0:33–0:34
A small pixelated window sits centered on a black field, showing a low-res 3D corridor with a sandy wall and a door; the view inside pans as the camera moves forward. A white monospace caption reads below the window.
On-screen text: “WorldModelGym”, “Take a trained agent.”, “test it directly.”
Camera: Static
21 · Demo · 0:34–0:36
A low-resolution pixelated 3D environment sits centered on a black field, first a narrow corridor, then a cut to a wider room marked with a red A, while captions appear along the bottom.
On-screen text: “Take a trained agent.”, “Run the exact same actions in the real environment and inside the world model.”
Camera: Static
22 · Demo · 0:36–0:37
A low-resolution voxel corridor plays in a centered window on a black field; the first-person view advances up the stone steps into a lit hallway. A two-line white monospaced caption sits below the window.
On-screen text: “Run the exact same actions in the real environment and inside the world model.”
Camera: Static
23 · Demo · 0:37–0:38
A pixelated first-person view of a torch-lit brick corridor, the camera rotating right so the far doorway drifts left, framed in a small window above a two-line white caption on a black field.
On-screen text: “Run the exact same actions in the real”, “environment and inside the world model.”
Camera: Static
24 · Demo · 0:38–0:39
A low-resolution 3D environment plays inside a centered frame on black. The camera pushes forward from a dim brick room through a doorway into a sunlit stone alley with a tree and an arch. A two-line white caption sits below the frame.
On-screen text: “Run the exact same actions in the real environment and inside the world model.”
Camera: Static
25 · Demo · 0:39–0:40
Two vertical panels sit side by side on black. The left is a blocky, RGB-glitched version of the scene; the right shows a clearer 3D game view of an armed figure in a desert town who leaps forward. White monospace captions swap at the bottom.
On-screen text: “Run the exact same actions in the real”, “environment and inside the world model.”, “Then compare what happens.”
Techniques: Split screen
Camera: Pull out
26 · Demo · 0:40–0:42
Two side-by-side vertical panels on black show the same third-person game scene in a sunlit courtyard, the left one pixelated and the right in full resolution as the character turns. A white caption sits below.
On-screen text: “Then compare what happens.”
Techniques: Split screen
Camera: Pull out
27 · Social proof · 0:42–0:45
A dim checkerboard of grayscale thumbnails brightens behind the centered white monospace line “Across 100+ tasks.” Four teal label chips (atari, environment, control, robotics) fade in at the corners, scale slightly, then fade out as the grid dims again.
On-screen text: “Across 100+ tasks.”, “atari”, “environment”, “control”, “robotics”
Techniques: Kinetic type
Camera: Pull out
28 · Other · 0:45–0:48
A white robotic hand holds a glossy red egg on its open palm against a dark backdrop. The egg lifts off the fingers and floats in the air above the hand while a five-line sentence fades in at the upper left.
On-screen text: “When the world model is faithful, its predictions are similar to the real world.”, “Across 100+ tasks.”
Techniques: 3D mockup
Camera: Static
29 · Other · 0:48–0:50
A white robotic hand, palm up, bounces a glossy red ball against a dark gradient background. The ball drops and rebounds while the hand tilts slightly. A block of white text sits on the left.
On-screen text: “When the world model is faithful, its predictions are similar to the real world.”
Techniques: 3D mockup
Camera: Static
30 · Other · 0:50–0:51
On a black field, teal dot-matrix particles form a crescent in the top left and a figure in the lower right. The crescent dissolves as the figure drifts to center and reshapes into a head profile.
On-screen text: “real world.”
Techniques: Particles, Morph
Camera: Push in
31 · Outro · 0:51–0:51
On black, two logos emerge from a soft blur: the white NEKA wordmark on the left and, on the right, a mark of two squares with a blue dot next to the gray word Labs. A faint dot grid shimmers at the bottom before the logos settle.
Camera: Static
32 · Outro · 0:51–0:54
Two white logo wordmarks, reka on the left and a four-square mark with Labs on the right, sharpen from a soft blur on a black field, then cut to a blue bar reading Read more.
On-screen text: “Read more”
Camera: Static
33 · CTA · 0:54–0:58
A solid cyan rectangle sits centered on a black field, carrying the dark monospace line Read more at reka.ai/labs. The block and text hold still, with no movement across the clip.
On-screen text: “Read more at reka.ai/labs”
Camera: Static
34 · CTA · 0:58–0:59
A cyan bar with dark text reading Read more at reka.ai/labs holds center on a black background, then the frame goes to black.
On-screen text: “Read more at reka.ai/labs”
Camera: Static



