
Wan 2.2 TI2V-5B: 5 s of 720p in under nine minutes on a consumer GPU
Wan 2.2's card says TI2V-5B makes 5 seconds of 720P video in under nine minutes on one consumer GPU. What that means for a queue, and a hosted comparison.
How Sume bills API work: plans and one wallet, what one run cost, estimates before you run, avatar video rates, failed jobs, and 402 errors.
Start with: How Sume pricing works: plans, one wallet, published model rates

Wan 2.2's card says TI2V-5B makes 5 seconds of 720P video in under nine minutes on one consumer GPU. What that means for a queue, and a hosted comparison.

A 250-character voiceover is $0.012 on Sume TTS. A 15-second clip is $1.88 on Wan 3.0 at 720p. Where the money goes in a short ad, with the arithmetic shown.
At 720p Seedance 2.5 is $0.578 per second on Sume and Wan 3.0 is $0.125. The gap comes from billing by video tokens (21,600 per second) versus a flat rate.

MiniMax's guide times one H3 run on 4 H200s at 75.1 s, about 53.7 s with Cache-DiT. The GPU-seconds, and the hourly rate where renting breaks even.

Two new-model promos, two dates. FLUX 3 Image's 50% launch price ended Oct 8; Vidu Q4's 30% image-to-video discount runs to Nov 30. Rates read Oct 9, 2026.

Luma's API lists a 5-second 540p video-to-video job at $0.72 and text-to-video at $0.15. The full table, and what a Gemini Omni edit costs on Sume.

Runway's API lists Gen-4 Turbo at 5 credits a second and Gen-4.5 at 12, with credits at $0.01. What 10 seconds cost, next to the Sume rates.

FLUX.2 Pro is $0.0375 per image on Sume and FLUX.2 Flex $0.0625. 2,000 stills: $75.00 or $125.00. FLUX 3 Image is $0.048 at 1K on fal but not in Sume's catalog.

Runway's API prices Veo 3.1 at 20 credits a second silent and 40 with audio, at $0.01 a credit. An 8-second example, and what audio costs on Sume.

A 25-second Seedance 2.5 trailer is $6.72 at 480p, $14.45 at 720p and $35.54 at 1080p on Sume. Fifteen trailers: $100.75, $216.68 or $533.08.

Inworld's pay-as-you-go TTS-2 is $25 per 1M characters, Flash $15. Sume TTS is $47.50. A year of 20 ads a month at 600 characters: $3.60, $2.16 or $6.84.

Google says Nano Banana 2.1 interim thought images are not charged. What that means for your invoice and how Sume bills one flat price per finished image.

A 30-second Wan 3.0 clip at 720p is $3.75 on Sume in one job; 40 listings are $150.00. The same set at 480p is $75.00 and at 1080p $300.00.

Vidu returns credits per task. Sume's poll response returns usage.cost in dollars on completed jobs. How to total spend and when the number appears.

Vidu Q4 image-to-video clips run 3 to 16 seconds. What one clip costs at 540p, 1080p and 4K with the 30% promo, and after November 30. Sume does not list it.

Omni Flash is $0.03 to $0.30 per second on fal; Sume bills list x 1.25. Price of the 8-second default, a 10-second max clip and a minute at each resolution.

Ideogram 4.5 is $0.03, $0.06 or $0.22 per image on fal by quality, size-independent. Sume bills list x 1.25: $0.0375, $0.075, $0.275. Per-1,000 table.

Sume prices a Seedance 2.5 job with a reference video at 60% of output plus an assumed 15 s of input. 5 s clips cost 2.4x; 30 s clips cost less. Table at 720p.

Google lists the Nano Banana 2.1 free tier as Not available. A look at paid per-image prices, what Google's figure leaves out, and Sume's flat billed price.

A weekly 15-second Wan 3.0 clip at 1080p is $3.75 on Sume, $195.00 for 52 weeks. At 720p $97.50, at 480p $48.75. With one retake in four, add 25%.

Wan 3.0 is $0.05, $0.10 and $0.20 per second at 480p, 720p and 1080p on fal. Sume bills list x 1.25. Per-clip prices at 5, 15 and 30 seconds.

250 six-second Seedance 2.5 product clips at 480p are $403.02 on Sume ($1.612 each); at 720p $866.70. Token math from fal's rate, with the per-clip table.

How to verify a Sume video price from the vendor page: take the fal rate, multiply by 1.25. Worked for Wan 3.0, Gemini Omni Flash 1.1 and Seedance 2.5.

Luma's API prices HDR at twice SDR and HDR plus 16-bit EXR at three times. A 5-second 1080p example, and what Sume's MP4 pipeline does and does not keep.

Sume bills provider list x 1.25. On Wan 3.0 at 720p that is $12.50 vs $10.00 for 100 seconds and $1,250 vs $1,000 for 10,000. What the gap is, sized.

Seedance 2.5 bills video tokens, not seconds. Width x height x seconds x 24 / 1024, then the per-1,000 rate. A 20-line Python calculator, checked against fal.

A 6-slide image carousel plus a 15-second Wan 3.0 reel with voice, music and render costs $2.448 on Sume list prices. Line items, swaps and a monthly total.

One 38-second ad from 6 stills, five 8-second Wan clips, a 640-character voiceover, a music bed and a render costs $5.8554, or $0.1541 per finished second.

Cutting 100 clips into 20 five-clip ads costs $7.60 in small Sume jobs (trim, filter, detach, concat, compose, render) against $75.00 of footage.
A new Sume avatar costs $0.95 once. Spread over 15-second standard clips at $2.76 each, it adds 24 cents per clip at 4 clips and under 4 cents at 25.

Wan 3.0 at 720p plus Sume Video Upscale at $0.009/s costs 53.6% of native 1080p at 12 s: $1.608 vs $3.00. The table covers 8, 12 and 30 seconds.

LTX lists 2.3 and 2.5 side by side. 720p triples (Fast $0.03 to $0.09, Pro $0.04 to $0.12); 4K rises only 22-25%. Table and 10-second math.

The minimum duration differs by model: Wan 3.0 starts at 2 s, Omni Flash at 3 s, Seedance at 4 s, MiniMax H3 at 5 s. Cost of one minimum clip at two sizes each.

Google's per-10-second price implies 25 audio tokens a second, so a minute of Gemini 3.8 Flash TTS costs 1.35 cents in 2026. Compared with Sume per minute.

Google lists Pro at $0.134 and 2.1 at $0.0336 for 1K (about 4x). Sume bills Pro $0.1875 and 2.1 $0.15 at 2K, only 1.25x. The ratio at each size.

On Sume, GPT Image 2.5 at low 1K lists at $0.02475 per image, below the $0.025 of Grok Image, Imagen 4 Fast and Qwen Image. 1,000 drafts: $24.75 against $25.00.

Sume text to speech costs $0.0475 per 1,000 characters and one request caps at 20,000 characters ($0.95). A 47,000-character chapter needs three requests.

Music 1.0 is a flat $0.125 per generation, with no duration parameter. 63 videos on 9 shared briefs cost $1.125; three takes per brief cost $3.375.

A 42-second talking still costs $4.3304 with MiniMax H3 Max Lip Sync (768p) and $8.0054 with VEED Fabric (720p), with the still and 640 characters of TTS.

Sume bills the seconds you generate, not the seconds you keep. Asking Wan 3.0 for 30 s and using 12 costs $3.75 for 12 s of use, $0.3125 per kept second.

BFL lists FLUX.2 Flex from $0.05 for typography work. Sume lists flux-2-flex at $0.0625 per image, so 24 posters cost $1.50. What the 'from' price leaves out.

MiniMax bills H3 Max reference video at $0.143 per second at 768P, on top of output. A 10-second clip with 5 seconds of reference is $1.515 direct.

Ray 3.2 single-keyframe extend is $0.15, $0.30 or $1.20 per call at 540p, 720p, 1080p, even if duration is set. Chain math and Sume's per-clip equivalents.

A 7-second 720p clip: $0.56 on MiniMax, $0.63 on LTX Fast, $0.525 on Sume H3. Luma's page lists only 5 and 10 second totals ($0.30, $0.90).

On LTX's price page Pro is $0.12-$0.39 per second against Fast at $0.09-$0.30. That is +30 to +33% at each tier; per-minute math and the nearest Sume rows.

What 30 seconds of 1080p costs: Luma as six 5-second clips ($7.20) or three 10-second ones ($10.80), LTX Fast $3.90, and Sume. Arithmetic shown.

Ray 3.2 edits bill $0.54 per 5 seconds at 360p against $0.06 for generation. The gap shrinks to 1.8x at 1080p. Table, test-loop math, and what Sume lists.

Vidu's table runs 9 to 78 credits per second, 540p to 4K. The two published USD points ($0.045, $0.39) imply half a cent per credit. Plus Sume's 4K row.