CumulusCumulus Labs
LeaderboardCategories
PlaygroundLogin
Get Started
LeaderboardCategoriesPlayground
Get Started
Back to Leaderboard
Video

FramePack I2V

+18.3%

~$FPACK

Frame-packing image-to-video model. Per-frame conditioning for highly consistent motion. Efficient architecture enables longer video sequences than comparable models.

Deploy Now
#22 overall · 1 deploys
14.2s
Cold Start
avg on Ion
120.0s
Avg Inference
per request
1
Active Replicas
right now
1
Total Deployments
all time

7-Day Trend

+18.3% this week
MonTueWedThuFriSatSun

Cold Start vs Competitors

Lower is better
●~$FPACK14.2s
~$FW1416.8s
~$FWTI512.4s
~$FW1-38.2s

Cold start comparison vs similar models. Lower is better.

Cost Estimate

Cost per 1K Inferences
$28.00
Est. Daily (1K req/day)
$28.00
Est. Monthly (30K req)
$840.00

No subscriptions. Buy credits, pay per inference. Scale to zero when idle.

Quick Deploy

cumulus-sdk
import cumulus from "cumulus-sdk"

// Deploy FramePack I2V on Ion
const client = await cumulus.deploy("framepack-i2v")

// Run inference
const result = await client.run({
  prompt: "Your prompt here",
  // model-specific params...
})

More in Video

FastWan T2V 14B~$FW14
16.8s+42.1%
FastWan TI2V 5B~$FWTI5
12.4s+38.6%
FastWan T2V 1.3B~$FW1-3
8.2s+35.2%
Wan 2.2 T2V A14B~$WAN14
18.5s+32.4%
View all Video models →
← Iondex

© 2026 Cumulus Compute Labs Corporation