LTX Studio LTXV: Lightricks’ Open-Source AI Video Model Hits 60-Second Clips

Date: July 12, 2026 | Reading Time: 7 minutes

What Is LTX Studio and LTXV?

Lightricks, the company behind the popular Facetune and Lightroom-style editing apps, entered the AI video space with LTX Studio — a full filmmaking environment that wraps AI video generation into a production workflow. At the core of LTX Studio is LTXV (also called LTX Video), an open-source AI video generation model that turns text prompts or still images into video clips.

Unlike closed-source models from Runway or Pika, LTXV is available as an open-weight model that creators can download and run locally, modify, fine-tune, and integrate into custom workflows. The model supports multiple variants — 2B parameters (faster, lower quality) and 13B parameters (higher quality, more VRAM) — with distilled versions of each that trade some fidelity for significantly faster generation.

For video creators evaluating AI tools, LTXV occupies a specific niche: it is not the highest-quality AI video model on the market (Runway Gen-4 Alpha holds that title), and it is not the fastest real-time model (Decart MirageLSD holds that title). But it is the most practical open-weight model for creators who want local, private, customizable AI video generation without per-query API costs.

The 60-Second Generation Capability

LTXV’s progression to 60-second clips happened through updates rather than a single launch. The original LTX Video model generated approximately 5 seconds of high-quality video in about 4 seconds on consumer hardware — impressive for a local model. A later autoregressive update in July 2025 extended continuous generation to approximately 60 seconds.

The 60-second capability works through autoregressive extension: the model generates an initial clip, then extends it by conditioning new frames on the previously generated output. This is different from native 60-second generation (a single pass that produces a full minute of video), but the practical result is the same — creators can generate full-minute clips from a single prompt or image without stitching together shorter segments.

This matters for video creators because most AI video models cap out at 5-30 seconds per generation. Producing a 60-second clip previously required generating multiple short clips, editing them together, and managing transitions between segments. LTXV’s autoregressive approach produces a single continuous output that does not need stitching, reducing the post-production overhead for longer-form AI-generated content.

LTX Studio: More Than a Model

LTXV powers LTX Studio, Lightricks’ broader AI video production environment. Unlike standalone models that generate clips and stop, LTX Studio provides a complete workflow: scripting, storyboarding, timeline editing, character consistency, camera control, voiceover, and collaborative production tools that wrap the AI generation into a filmmaking pipeline.

Key features that differentiate LTX Studio from other AI video tools:

LoRA-based conditioning. Creators can fine-tune LTXV’s output with Low-Rank Adaptation (LoRA) models that train specific styles, character appearances, or visual treatments. A creator can train a LoRA on their brand’s visual style and apply it consistently across all generated content without re-prompting every generation.

Multiple model variants. The 2B model runs on consumer GPUs with approximately 4-6 GB VRAM — accessible to most creators with mid-range hardware. The 13B model requires 12-16 GB VRAM but produces noticeably better output quality. Distilled variants offer intermediate tradeoffs. This flexibility lets creators choose the right tool for each project rather than being locked into a single quality/performance level.

Full production pipeline. LTX Studio includes script-to-video, storyboard-to-video, and timeline editing — not just single clip generation. Creators can build multi-scene projects within the tool rather than exporting clips to a separate video editor.

Open-weight access. Because LTXV is open-weight, creators are not tied to a subscription or API credits. They download the model once and generate as much as their hardware can handle. For heavy users, this economics advantage over per-generation pricing models is substantial.

How LTXV Compares to Other AI Video Models

Runway Gen-4 Alpha — Higher quality output than LTXV, better cinematic fidelity, more sophisticated camera control. But closed-source, subscription-based ($15-95/month), and limited to shorter clips. Best for: marketing content where quality matters more than cost or control.

Pika Director’s Suite — Stronger narrative and character consistency across scenes. Closed-source with credit-based pricing ($10-70/month). No local running option. Best for: story-driven content series where character persistence is critical.

Decart MirageLSD — Real-time live-stream transformation, not clip generation. Sub-40ms latency for continuous streaming. Requires H100 GPUs, not consumer hardware. Best for: live streaming and real-time video transformation, not clip creation.

Meta Muse — Long-form character-driven generation with high consistency across extended narratives. Closed-source, cloud-only. Best for: episodic content where character and setting consistency across episodes is paramount.

LTX Studio / LTXV — Open-weight, local-running, fast generation on consumer hardware. 60-second autoregressive clips. LoRA support for custom styles. Full production workflow. Best for: creators who need local, private, customizable AI video generation without per-query costs, and who want a complete production environment rather than a single-clip tool.

Why Creators Should Pay Attention to Open-Source AI Video

The emergence of capable open-weight models like LTXV is the most important trend in AI video for independent creators. Closed-source models charge per generation or require ongoing subscriptions — a model that costs $15-95/month for moderate use, plus additional credits for heavy generation, creates a variable cost that scales with content output.

An open-weight model that runs locally has a fixed cost: the hardware you already own. A creator generating 100 videos per month with a closed-source tool could pay $50-200/month in subscription and credit costs. The same creator running LTXV locally pays $0 in per-generation fees — just the electricity and GPU depreciation.

The tradeoff is quality and convenience. LTXV’s output is not as polished as Runway or Pika. Setting up a local model requires some technical comfort. But for creators who generate high volumes of content and cannot justify per-clip costs scaling with their output, the open-weight path is increasingly viable.

LTXV’s support for LoRA training adds another dimension: creators can train custom styles, character appearances, and visual treatments that persist across all generations. A closed-source model may offer style transfer or reference image features, but an open-weight model with LoRA gives the creator full ownership of the visual identity — no reliance on a provider’s evolving feature set.

Real Workflows: When To Use LTXV vs Other AI Video Tools

Choosing the right AI video tool depends on what you are creating and how you need to produce it. Here is a practical decision framework based on common creator workflows:

Social media short-form content (TikTok, Reels, Shorts). These clips are typically 15-60 seconds. LTXV covers the full length in a single generation. For quick, iterative content where you generate multiple variations and pick the best, the zero marginal cost of local generation (no API credits) is a significant advantage over per-clip pricing models. Quality gap with Runway is noticeable in close-up shots and complex motion, but acceptable for fast-paced social content where viewers scroll past in seconds.

Brand marketing videos with specific visual identity. If your brand has a defined visual style — colors, lighting, aesthetic — LTXV’s LoRA support lets you train a custom model on your brand’s look. This is not possible with closed-source tools that offer generic style presets. A brand that generates consistent visual content at scale benefits from having their entire visual identity encoded in a LoRA model rather than re-prompting for consistency every time.

Client-facing content where quality is paramount. For high-stakes client deliverables (commercials, product launches, premium content), Runway Gen-4 Alpha’s superior output quality justifies the subscription cost. LTXV is not there yet on pure quality. Runway’s camera control and cinematic output produce results that pass for professional production; LTXV results carry a clear “AI-generated” visual signature.

Bulk content generation for editorial or evergreen libraries. A content creator producing 50-100 videos per month for YouTube, educational platforms, or automated content pipelines will find LTXV’s economics transformative. At $15-95/month for Runway plus additional credits, 100 videos could cost hundreds of dollars in a closed ecosystem. LTXV runs on hardware you already own with zero per-video cost. The quality tradeoff is acceptable for content where the message matters more than the polish.

Experimental and iterative creative work. When exploring visual concepts, testing narrative structures, or iterating on creative directions, the ability to generate dozens of variations without incremental cost changes how creators approach the creative process. With per-generation pricing, each iteration costs money, discouraging experimentation. With local generation, the only cost is time.