1:25
Belvedir Launch
Belvedir · B2B · Aug 2026
Explore
Subquadratic — Efficiency is Intelligence
Subquadratic · Nov 2026 · 1:16 · AI-annotated
- Length
- 1:16
- Scenes
- 28 · 2.7 s avg
- Product on screen
- 0:06
- Features shown
- 1
- Call to action
- 1:11 · 93%
- Voiceover
- Yes
00:00:00 / 01:16:08
Subquadratic’s 1:16 launch film runs 28 scenes, about 2.7 s each. The product shows up at 0:06 and 1 feature is revealed, with social proof. The call to action lands at 1:11 (93% in), over a voiceover. The video alternates locked-off talking-head shots in a beige office with flat grid-paper graphics, cutting at a brisk ~19 cuts per minute. Its signature moves are pixel-grid wipes, sliding-bar text reveals, and stat bars that tick up in sync with the voiceover.
Footage neededEstimated needs: 1 capture group — your screens or recordings · 1 official logo — reused across the film · 12 filmed passages — shoot, supply or replace · 5 graphic passages — draw or supply
Is this your launch? to your site or press kit.
More like this
Subquadratic — Efficiency is Intelligence, scene by scene
· 28 scenes as text
01 · Hook · 0:00–0:06
A man in a gray t-shirt stands against a beige wall as a colorful pixel grid clears to reveal him. Colored squares over his chest give way to the white serif headline, revealed behind sliding bars, while a caption sits below.
On-screen text: “Today, we're announcing the”, “The First”, “Long-Context Decision Model”
Techniques: Masking, Kinetic type
Camera: Static
02 · Product intro · 0:06–0:10
A man in a gray tee speaks to camera in a bright office, blurred city skyline behind him. Captions run along the bottom as the SubQSnap logo with green corner brackets fades in at left. The shot cuts to a dark grid background where the serif title scales in.
On-screen text: “Powered by Subquadratic Sparse Attention,”, “SubQ Snap's multimodal context window”, “SubQSnap”, “The First”, “Long-Context Decision Model”
Techniques: Kinetic type
Camera: Static
03 · Product intro · 0:10–0:11
On a black grid, a large white serif 1 expands into 12 as a yellow serif X stamps in to the right. A small white caption below reads is 12 times larger than, then swaps to Jev's.
On-screen text: “1”, “is 12 times larger than”, “12X”, “Jev's.”
Techniques: Kinetic type, Masking
Camera: Static
04 · Social proof · 0:11–0:16
A beige slide holds a benchmark table under the title Decision Model Benchmarks, a green highlight stepping down the SubQSnap column row by row. It then cuts to a man in a gray t-shirt by a window, speaking, with a caption at the bottom.
On-screen text: “Decision Model Benchmarks”, “It isn't just better.”
Camera: Static
05 · Problem · 0:16–0:19
A man in a gray t-shirt stands beside a window, a blurred city skyline behind him, and speaks to an off-camera interviewer. A dark caption bar at the bottom displays his words in white text. The camera holds a locked-off medium close-up.
On-screen text: “It isn't just better.”, “It is a different type of decision model.”
Camera: Static
06 · Problem · 0:19–0:20
A man in a gray t-shirt stands centered against a beige wall, speaking to camera in a locked-off medium shot, with a leafy plant at the right edge. A white caption on a dark rounded pill sits at the bottom center.
On-screen text: “It is a different type of decision model.”, “Here's why.”
Camera: Static
07 · Problem · 0:20–0:24
On a dark grid, a square box labeled Frontier LLM holds a patchwork of colored blocks that drift rightward into a second box labeled Decision model, which fills with color. The shot then cuts to a man in a gray tee against a window, locked off.
On-screen text: “Frontier LLM”, “Decision model”, “Here's why.”
Camera: Pull out
08 · Problem · 0:24–0:27
A man in a gray t-shirt talks to camera in a soft-lit room, a window with blinds blurred behind him. Lower left, white serif text 8K-32K Context Window scales in, then a thin red line strikes through both lines and the text dims to gray.
On-screen text: “8K-32K”, “Context Window”
Techniques: Kinetic type
Camera: Static
09 · Problem · 0:27–0:30
A man in a gray t-shirt stands centered against a beige wall, a leafy plant at the right edge, speaking to camera in a locked-off medium shot. White caption chips read his words at the bottom as a larger 8K-32K label scales in.
On-screen text: “8K-32K”, “simply can't handle”, “important decisions for enterprises,”
Camera: Static
10 · Problem · 0:30–0:34
A man in a gray t-shirt sits in a softly lit room beside a window with blinds, speaking toward an off-camera interviewer. The shot is a static medium close-up. White subtitles on a dark rounded bar appear at the bottom of the frame as he talks.
On-screen text: “important decisions for enterprises,”, “with multi-million line”, “codebases or knowledge”, “bases of hundred-page documents.”
Camera: Static
11 · Problem · 0:34–0:36
A man in a gray t-shirt stands centered against a pale green wall, a leafy plant at the right edge, speaking directly to camera. A subtitle line sits in a dark rounded bar at the bottom of the frame, static throughout the locked-off shot.
On-screen text: “bases of hundred-page documents.”, “That's what Snap was trained for.”
Camera: Static
12 · Feature reveal · 0:36–0:38
A cream field holds a horizontal orange bar and gray bar that wipe clear, then the serif word Agentic types out letter by letter to read Agentic Decisions.
On-screen text: “That's what Snap was trained for.”, “Agentic Decisions”
Techniques: Kinetic type, Masking
Camera: Static
13 · Feature reveal · 0:38–0:41
A man in a gray t-shirt talks to camera in a medium close-up, placed right of a sunlit window with blinds. White subtitles caption his speech along the bottom. The framing then cuts to a centered medium shot against a plain wall, a leafy plant at right.
On-screen text: “Agentic Decisions”, “where it is purpose-built to tackle”, “long-horizon agentic workflows.”
Camera: Static
14 · Feature reveal · 0:41–0:44
A man in a gray t-shirt sits centered before a pale olive wall, a leafy plant at the right edge, speaking directly to camera. A dark rounded caption bar with white text sits at the bottom, swapping lines as he talks. The frame is locked off and static throughout.
On-screen text: “long-horizon agentic workflows.”, “It is able to dramatically reduce”, “cost and latency”
Camera: Static
15 · Feature reveal · 0:44–0:46
A flat document-analysis UI mockup sits right of a left column holding a serif headline and bold stats. The card's page-grid lights up, then swaps to a red No answer with a cited clause and a stats row. A dark pill caption cycles phrases along the bottom.
On-screen text: “cost and latency”, “within tasks like:”, “Finding insights and contradictions”
Camera: Static
16 · Feature reveal · 0:46–0:48
A flat two-column layout: left, a serif document title above three large sans statistics; right, a policy-check panel showing a claim, a red 'No' verdict card, and a quoted clause, with a stats row beneath. The composition holds still while a dark caption pill at the bottom swaps through two phrases.
On-screen text: “Finding insights and contradictions”, “in large documents”
Camera: Static
17 · Feature reveal · 0:48–0:51
A flat lodging search interface fills the frame with a dense grid of listing cards. A cursor clicks the Search again button, and the grid reflows into a single row of four numbered cards. A dark caption box sits at the bottom.
On-screen text: “Building a multimodal”, “recommendation system”
Techniques: Cursor demo
Camera: Static
18 · Feature reveal · 0:51–0:52
A dark interface shows a grid of file cards, most bordered in red with crossed-out labels, while several turn green with checkmarks as a timer ticks up. A cursor rests on the Running button, and a white caption sits centered at the bottom.
On-screen text: “Curating context within coding agents”
Techniques: Cursor demo
Camera: Static
19 · Feature reveal · 0:52–0:53
A dark interface view builds up: four file rows with green probability tags and segmented range bars fade in on the left, a right panel reads 7 kB and 4 files, a timer ticks up top right, and a white caption sits at the bottom.
On-screen text: “Curating context within coding agents”
Camera: Static
20 · Feature reveal · 0:53–0:54
A static dark infographic shows the headline The agent starts from this packet, not the whole repo above four file rows with green coverage bars, an arrow, and a blue 7 kB card. The bottom caption fades out.
On-screen text: “The agent starts from this packet, not the whole repo”, “Curating context within coding agents”
Camera: Static
21 · Feature reveal · 0:54–0:55
A flat UI panel stays locked on a hot dog photo under a green “Hot dog” banner with a checkmark. A bar reads 98% sure, 219 ms, 431 tokens; a right column lists food thumbnails. A bottom caption reads “Or deciding if an item is a hot dog.”
On-screen text: “Or deciding if an item is a hot dog”
Camera: Static
22 · Feature reveal · 0:55–0:56
A flat UI mockup of a classification app: a pizza photo with an orange banner and a caption pill, while a confidence bar fills and stats tick up. A pointer cursor hovers over a pizza thumbnail in a right-hand image column.
On-screen text: “or is not a hot dog”
Techniques: Cursor demo
Camera: Static
23 · Feature reveal · 0:56–1:03
A grid-paper infographic compares two side-by-side panels, LLM and SubQ Snap, as workflow steps fade in with orange markers and green Cost and Latency bars grow with ticking values. A dark caption pill at the bottom swaps phrases in sync with the voiceover.
On-screen text: “Snap does this by”, “taking the data-intensive”, “and structured decisions,”, “while the LLM handles”, “workflow steps that are open-ended.”
Camera: Static
24 · Social proof · 1:03–1:05
A man in a gray t-shirt stands in a softly lit room beside a plant, speaking to camera as white subtitles appear on dark pills at the bottom. The shot cuts to a cream title card reading Long-Context Reasoning with a small data row of percentages and labels.
On-screen text: “It can even one-shot long-context”, “problems on its own, like temporal”, “Long-Context Reasoning”, “Decision LCR - 100K tokens per question”, “0.0%”, “0.0%”, “SubQSnap”, “Jev”, “one-shot”, “*chunked”
Camera: Static
25 · Social proof · 1:05–1:08
On a beige slide, a bar chart titled Long-Context Reasoning tweens from three bars to two: the green bar drops slightly, the middle gray bar shrinks away, and the right gray bar grows to 48.0% as the labels slide into place.
On-screen text: “Long-Context Reasoning”, “Decision LCR - 100K tokens per question”, “problems on its own, like temporal”
Camera: Static
26 · Other · 1:08–1:11
A man in a gray t-shirt stands centered against a pale green wall, speaking to camera in a locked-off medium shot. A leafy plant fills the right edge and a window sits left. White subtitles in a dark box appear line by line at the bottom.
On-screen text: “This represents”, “a fundamental shift to the AI”, “stack for enterprises.”
Camera: Static
27 · CTA · 1:11–1:13
A black frame holds the small still line “stack for enterprises.” A red bar and a blue block wipe across the center, then the white serif text “Visit Subq.ai” scales up in the middle and holds.
On-screen text: “stack for enterprises.”, “Visit Subq.ai”
Techniques: Masking, Kinetic type
Camera: Static
28 · Outro · 1:13–1:16
The Subquadratic logo holds static on a black background: a small grid of colored squares (yellow, blue, red, green with a gray center) beside the white serif wordmark. The frame is locked-off with no motion.
Camera: Static