Cinematic stage
Warm gold spotlight stage loop for LTX 2.5 hero mood.
Hot news · LTX 2.5
LTX 2.5Is Here
Faster AI video, native ComfyUI, production-ready pricing
LTX has launched LTX-2.5 — an ultra-fast video generation model with native ComfyUI integration. Generate a 10-second 720p clip in just 6.8 seconds on two NVIDIA GB200 GPUs. LTX-2.5 Fast delivers audio-ready 720p from $0.09 per second, and eligible organizations can use it free.
Sample reels on this page were generated with local LTX video models and are AI-generated.
LTX 2.5 at a Glance
Headline numbers from the LTX-2.5 launch
6.8s
For 10s 720p video
$0.09/s
LTX-2.5 Fast rate
720p
With synced audio
<$10M
ARR free eligibility
LTX sample reels
Generated on this machine with LTX-Video, compressed, then hosted on CDN44 R2 — not shipped in app/public.
Golden-hour coast
Aerial glide over cliffs — continuous single-take motion.
Realtime frames
Abstract panel of cascading frames — speed metaphor.
Why LTX 2.5 Matters
Realtime-class speed
On a dual NVIDIA GB200 setup, LTX-2.5 generates a full 10-second 720p video in about 6.8 seconds — faster than real time for many production loops.
Native ComfyUI integration
LTX-2.5 ships with first-class ComfyUI support so teams can drop the model into existing node graphs without fragile custom wrappers.
Transparent Fast pricing
LTX-2.5 Fast prices audio-ready 720p at $0.09 per second of output. A 10-second clip costs about $0.90 — easy to forecast for batch jobs.
Free for growing teams
Organizations with annual recurring revenue under $10 million can use LTX-2.5 free under the published eligibility rules.
720p video with audio
LTX-2.5 Fast targets production clips that include audio — not silent previews — so drafts are closer to shippable deliverables.
Built for workflows
From prompt-to-preview loops to ComfyUI pipelines, LTX 2.5 is positioned as a practical upgrade for creators, studios, and product teams.
Benchmark Snapshot
Published dual-GB200 reference configuration
Generation time
6.8 seconds
To produce a 10-second clip
Reference hardware
2× NVIDIA GB200
Dual-GPU launch config
Output target
10s · 720p
With audio on Fast tier
Figures reflect publicly reported LTX-2.5 launch benchmarks. Actual cloud latency depends on queue, resolution, and length.
LTX-2.5 Fast Pricing
Simple per-second economics for audio-ready 720p
LTX-2.5 Fast
Pay per second
$0.09
per second of 720p video with audio
- 10-second clip ≈ $0.90
- Audio-ready 720p output
- Predictable cost for batch generation
Eligibility
Free for qualified orgs
$0
when ARR is under $10 million
- Organizations under $10M ARR
- Use LTX-2.5 under published free terms
- Ideal for startups and small studios
Native ComfyUI Support
LTX 2.5 is designed to plug into ComfyUI out of the box. Creators who already chain loaders, samplers, and post nodes can adopt LTX-2.5 without rebuilding their stack.
- First-class ComfyUI nodes and workflows
- Fits existing AI video pipelines
- Faster iteration for local and remote setups
Workflow-ready model
Whether you run cloud generation on LTX-2 AI or local ComfyUI graphs, LTX-2.5 is the release aimed at both product UX and technical power users.
LTX 2.5 FAQ
LTX 2.5 (LTX-2.5) is the latest LTX AI video model release, focused on generation speed, native ComfyUI integration, and clearer commercial pricing via LTX-2.5 Fast.
Start with LTX 2.5 today
Generate fast 720p AI video with audio — online on LTX-2 AI
Open Video Studio Ultra-fast generation 720p + audio path ComfyUI ready