Reviewed by Jonathan West · Updated Sep 9, 2026

LTX-2.5 Pricing: Per-Second Rates, Studio Plans, and Open Weights

A breakdown of Lightricks LTX-2.5 costs across per-second generation billing, LTX Studio subscription tiers, and open-source licensing thresholds.

Reviewed by Jonathan West · Updated Sep 9, 2026

LTX-2.5 pricing operates across three distinct distribution tiers, including free open weights for businesses under ten million dollars in annual revenue, managed application programming interface (API) access billed per second of output, and creative application subscriptions.

Lightricks introduced LTX-2.5 in October 2026 as an open 22B-parameter diffusion model that supports both text-to-video and image-to-video generation across resolutions from 720p up to 4K.

Selecting the most cost-effective path depends on output volume, infrastructure hosting capabilities, and whether your organization requires commercial distribution rights.


LTX-2.5 Pricing at a Glance

LTX-2.5 pricing offers three distinct access models comprising open weights, usage-based cloud API generation, and packaged web application tiers. Lightricks distributes the 22B-parameter model weights at no software charge for commercial organizations with under ten million dollars in revenue. Production environments requiring managed infrastructure pay per second of generated video through the official API, while visual creators can access the model via tiered web plans.

API usage carries no base request fees or per-asset surcharges on the Lightricks API pricing page. Billing accumulates strictly on output duration, with identical per-second rates applied to landscape and portrait aspect ratios across both text-to-video and image-to-video endpoints. Production teams can choose between the Fast variant for rapid iterations and the Pro variant for final renders across four standard resolutions.

  • LTX-2.5 Fast at 1280x720: $0.09 per second of generated video
  • LTX-2.5 Fast at 1920x1080: $0.13 per second of generated video
  • LTX-2.5 Fast at 2560x1440: $0.19 per second of generated video
  • LTX-2.5 Fast at 3840x2160: $0.30 per second of generated video
  • LTX-2.5 Pro at 1280x720: $0.12 per second of generated video
  • LTX-2.5 Pro at 1920x1080: $0.17 per second of generated video
  • LTX-2.5 Pro at 2560x1440: $0.25 per second of generated video
  • LTX-2.5 Pro at 3840x2160: $0.39 per second of generated video

Fast and Pro Tier Cost Comparisons

The pricing spread between LTX-2.5 Fast and LTX-2.5 Pro reflects distinct computational profiles designed for different stages of the production pipeline. Lightricks created the Fast tier for rapid creative exploration, preview screening, and early concept iteration. In contrast, Lightricks tuned the Pro tier for higher fidelity, increased temporal stability, and production-ready output. At 1080p resolution, LTX-2.5 Pro costs seventeen cents per second compared to thirteen cents per second for LTX-2.5 Fast, which represents a cost increase of about thirty-one percent.

Budget calculations for single clips demonstrate how output resolution and model tier influence individual job costs. Based on arithmetic from the published rates, rendering a single ten-second clip at 1080p on LTX-2.5 Pro costs one dollar and seventy cents (ten seconds multiplied by seventeen cents). Rendering that same ten-second 1080p clip on LTX-2.5 Fast costs one dollar and thirty cents. A draft ten-second clip at 720p on LTX-2.5 Fast bills at ninety cents, while a high-resolution ten-second clip at 4K on LTX-2.5 Pro totals three dollars and ninety cents.

Volume generation pipelines accumulate noticeable cost variances when scaling up batch jobs. Generating one hundred ten-second draft clips at 1080p on LTX-2.5 Fast totals one thousand seconds of video, which bills at one hundred thirty dollars based on published rates. Producing a continuous sixty-second sequence on LTX-2.5 Pro at 1080p reaches ten dollars and twenty cents. Readers exploring the underlying architectural advancements can read our review of LTX-2.5 explained.


Price Differences Between LTX-2.5 and LTX-2.3

LTX-2.5 API rates represent an upward pricing adjustment compared to earlier LTX-2.3 generation tiers across all resolutions on the Lightricks pricing documentation. At standard 1080p resolution, LTX-2.5 Pro bills at seventeen cents per second compared to eight cents per second for LTX-2.3 Pro. This nine-cent difference means LTX-2.5 Pro costs slightly more than double the prior model generation rate.

The entry-level Fast tier displays a similar price increase at standard high-definition resolution. LTX-2.5 Fast at 1080p costs thirteen cents per second compared to six cents per second for LTX-2.3 Fast, representing an increase of seven cents per second. At 4K resolution, the proportional margin between generations narrows slightly: LTX-2.5 Pro costs thirty-nine cents per second versus thirty-two cents for LTX-2.3 Pro, while LTX-2.5 Fast costs thirty cents per second versus twenty-four cents for LTX-2.3 Fast.

Automatic duration billing mechanisms charge accounts based on the actual duration of completed video rather than a fixed estimate. When developers initiate jobs with automatic duration, prepaid accounts place a temporary credit hold against the maximum duration permitted for the specified resolution and frame rate. Once video processing finishes, Lightricks bills only the final rendered seconds and releases the remaining held credits back to the account balance.


Audio-to-Video and Specialized Tool Rates

Audio-to-video processing on LTX-2.5 bills directly on the duration of the input audio track at standard video generation rates. Lightricks accepts input audio in WAV, MP3, M4A, and OGG formats alongside optional PNG, JPEG, or WEBP reference image files. Individual audio-to-video requests process clips up to approximately twenty seconds in length, allowing developers to construct longer sequences by chaining successive requests together.

Specialized video modification endpoints operate under independent rate structures and older model dependencies on the Lightricks model page. Video editing through the Retake endpoint requires the older ltx-2-3-pro model, billing at ten cents per second of input video at 1080p resolution. Lightricks has not published an LTX-2.5 rate for Retake, nor does Lightricks publish rates for extend, reframe, or High Dynamic Range (HDR) conversions on LTX-2.5.


LTX Studio Subscription Tiers

Lightricks provides managed creative application access through tiered LTX Studio pricing plans that package generation credits with browser-based editing tools. The entry Free tier provides a one-time grant of 800 credits covering both LTX-2.5 and LTX-2.3 models. Lightricks restricts output generated on the Free tier to personal projects, explicitly prohibiting any use for marketing, business, or paid commercial work.

Paid subscription tiers unlock ongoing monthly credit allocations and expanded commercial licensing rights. The Lite plan costs fifteen dollars per month on monthly billing, or twelve dollars per month when billed annually at one hundred forty-four dollars per year, providing 8,000 monthly credits with editing tools, upscaling, and watermark removal. The Standard plan costs thirty-five dollars monthly, or twenty-eight dollars monthly billed annually at three hundred thirty-six dollars per year, supplying 28,000 credits and granting a commercial-use license for direct and indirect business projects.

Higher volume plans include the Pro tier at one hundred twenty-five dollars monthly for 110,000 credits, while Enterprise solutions require custom scoping through Lightricks sales representatives. Lightricks does not publish the precise number of credits consumed per second of LTX-2.5 generation, and temporary promotional discounts on the pricing page do not constitute permanent standing rates. Creative organizations evaluating collaborative video production can read our guide to LTX-2 for creative teams.


Open-Weights Licensing and Hardware Requirements

Lightricks distributes the complete 22B-parameter LTX-2.5 weights on Hugging Face alongside implementation source code on GitHub. On the main model overview page, Lightricks states that LTX-2.5 is free for companies with under ten million dollars in revenue. However, on the specific LTX-2.5 product page, Lightricks describes the qualification as free use under ten million dollars in annual recurring revenue (ARR).

The Hugging Face repository tags LTX-2.5 under an 'other' license category rather than a standard open-source definition. Because published statements alternate between total revenue and annual recurring revenue, technical leaders should confirm the exact license text on Hugging Face before deployment. Commercial organizations exceeding the ten million dollar threshold must negotiate a custom commercial license directly with Lightricks.

Self-hosting eliminates API fees but transfers all infrastructure costs directly to local hardware procurement or cloud graphics processing unit (GPU) rental. A comparison table on the LTX-2.5 product page lists a "Min VRAM: 16GB" row, but it is a marketing comparison, not a hardware specification, and no local inference speed figures are given in the sections reviewed, so verify on the vendor page and test on your own hardware. Developers estimating local hardware needs for self-hosted instances can consult our local AI hardware calculator.

The Hugging Face repository labels LTX-2.5 with an 'other' license tag, meaning qualifying organizations must verify the exact license terms directly before commercial deployment.

Selection Criteria for Deployment Paths

Selecting the optimal deployment path for LTX-2.5 requires balancing generation volume against infrastructure overhead and licensing constraints. Individual creators testing visual capabilities can evaluate the model at zero financial cost in the LTX Playground or via the 800 one-time credits on the LTX Studio Free plan. Within LTX Studio, commercial projects need the Standard plan or higher, since the Free plan is personal use only.

Managed API endpoints provide managed infrastructure for programmatic software products that run without dedicated GPU management. Self-hosting the open weights supports private deployment for high-volume pipelines, provided the business qualifies under the ten million dollar revenue limit and owns sufficient compute capacity. To cross-reference API costs across alternative video generators, see our directory of AI model pricing.

Before committing engineering resources, verify your team's throughput requirements against unpublished operational boundaries. Benchmark scores, a maximum clip length for the open weights and local inference speeds are not given in the vendor pages reviewed for this guide, so test a sample batch on your target infrastructure.

  • Choose LTX Studio Free or the Playground for zero-cost creative evaluation and prompt testing
  • Choose the managed API for programmatic application pipelines requiring automated per-second billing
  • Choose self-hosted open weights for high-volume generation, private deployment, or custom fine-tuning

What you need to run LTX-2.5 yourself

LTX-2.5 needs real memory, but it is within reach of a high-end workstation or a couple of professional GPUs — and many teams simply rent instead of buying. Match the path below to whether you want to own the hardware or pay by the hour.

PathWhat it isBest forGet started
Call the hosted APIUse LTX-2.5 as a pay-per-token API — zero hardwareMost teams; getting startedOpenRouter
Rent GPUs by the hourSpin up an H100 / A100 for a few dollars an hourFlexible self-hosting without buying cardsRunPod
Local on unified memoryOne Mac with enough unified memory to hold a 4-bit quantA single quiet on-prem boxApple Mac Studio (M4 Max, 128GB)
Local on a workstation GPUOne 48GB pro card, or two 24GB consumer cardsPower users who want hardware they ownNVIDIA RTX 6000 Ada (48GB)

To put LTX-2.5 to work once it is live, connect a coding client like Cursor (via OpenRouter) or a local runner such as Ollama.

Apple Mac Studio (M4 Max, 128GB)
Apple Mac Studio (M4 Max, 128GB)

A single quiet on-prem box

View on Amazon →
NVIDIA RTX 6000 Ada (48GB)
NVIDIA RTX 6000 Ada (48GB)

Power users who want hardware they own

View on Amazon →
Rule of thumb: a model needs roughly half its parameter count in gigabytes of memory at 4-bit — so a ~70B model wants about ~40GB. That fits one 48GB professional GPU, two 24GB consumer cards, or a 64–128GB unified-memory Mac. Below that budget, rent it by the hour instead of buying.

Frequently Asked Questions

  • LTX-2.5 API pricing ranges from nine cents to thirty-nine cents per second of generated video depending on resolution and model tier. The Fast variant costs $0.09 at 720p, $0.13 at 1080p, $0.19 at 1440p, and $0.30 at 4K. The Pro variant costs $0.12 at 720p, $0.17 at 1080p, $0.25 at 1440p, and $0.39 at 4K. Billing applies strictly to output duration with no request fees.
  • LTX-2.5 is free to download and self-host for companies under ten million dollars in revenue or annual recurring revenue (ARR). Lightricks also offers a Free tier on LTX Studio with 800 one-time credits for personal use only, and zero-cost testing in the LTX Playground. API generation and commercial LTX Studio subscriptions require paid fees.
  • LTX-2.5 Fast is designed for rapid iteration, previews, and creative exploration, while LTX-2.5 Pro is optimized for higher visual fidelity and temporal stability in production renders. At 1080p resolution, LTX-2.5 Pro costs seventeen cents per second compared to thirteen cents per second for LTX-2.5 Fast, making Pro roughly thirty-one percent more expensive.
  • A ten-second clip on the LTX-2.5 API costs between ninety cents and three dollars and ninety cents based on published rates. At 1080p resolution, a ten-second clip costs $1.30 on LTX-2.5 Fast and $1.70 on LTX-2.5 Pro. At 720p, a ten-second Fast clip costs $0.90, while a 4K Pro clip costs $3.90.
  • Yes, LTX-2.5 API pricing is higher than LTX-2.3 across all resolutions. At 1080p, LTX-2.5 Pro costs seventeen cents per second compared to eight cents for LTX-2.3 Pro, slightly more than double the prior rate. At 1080p Fast, LTX-2.5 costs thirteen cents per second compared to six cents for LTX-2.3 Fast.
  • Yes, commercial use is permitted under specific conditions. Companies under ten million dollars in revenue can use the open weights commercially without software licensing fees. For LTX Studio, commercial rights require the Standard plan ($35 per month) or higher, as the Free plan is strictly for personal use. Organizations exceeding the ten million dollar threshold must secure an enterprise commercial license from Lightricks.

Planning an LTX-2.5 Video Generation Architecture?

Book a consultation with Layer3Labs to evaluate API rate scaling, self-hosted hardware costs, and production integration trade-offs.

Book a Consultation
Disclosure: Layer3Labs is reader-supported. When you buy through links on this page we may earn an affiliate commission, at no extra cost to you.