AI models on Craze
Every video, image, voice, music and sound model Craze can direct, with what each one makes and what it costs.
Every video, image, voice, music and sound model Craze can direct, with what each one makes and what it costs.
Starts at$0.179 per second of audio
Lightricks' latest open-source video model for talking photos. Animate a still image with audio to create 1080p talking head videos.
LTX-2.5 syncs mouth movements to every phoneme.
More than just mouth movement.
Script the voiceover, have talent record it, upload a photo, and Craze generates the talking video.
No upscaling, no separate audio sync.
Craze creatively jams, researches and plans with you.
Craze generates anything imaginable. Video, audio, music, motion graphics and more.
Craze watches your footage and builds a final cut. With a real editor if you need it.
Every model here runs on every Craze plan, including Free, and Craze can use a different one for each shot.
| LTX 2.5Lightricks | Seedance 2.5ByteDance Seed Team | Seedance 2.0ByteDance Seed Team | Seedance 2.0 FastByteDance Seed Team | |
|---|---|---|---|---|
| Makes | Text + Audio + Image to video | Text + Image + Video + Audio to video | Text + Image + Video + Audio to video | Text + Image + Video + Audio to video |
| Quality | Up to 1080p | Up to 1080p | Up to 4K | Up to 720p |
| Length | — | Up to 30s per clip | Up to 15s per clip | Up to 15s per clip |
| Sound | — | Generated with the picture | Generated with the picture | Generated with the picture |
| From | $0.179 per second of audio | $0.114 per second of video | $0.075 per second of video | $0.06 per second of video |
| Task | Inputs | Resolution | Aspect ratios | Sound |
|---|---|---|---|---|
| Talking photo | Audio, First frame | 1080p | 16:9, 9:16 | None |
$0/ month
$19/ month
$59/ month
$199/ month
LTX-2.5 solves one problem perfectly: turning a still photo into a talking head video. When you have a script and a portrait, LTX-2.5 animates it with tight lip sync and natural facial movement. No avatar, no uncanny valley, just motion that fits the audio. It's built specifically for this task, not as a side feature of a general video model.
LTX-2.5 syncs mouth movements to every phoneme. The lip sync isn't an approximation, it's trained in. Watch the model work and you'll see mouths match dialogue, not speech that approximates the mouth. The animation understands the mechanics of speech, so /p/ sounds open the mouth differently than /m/ sounds.
More than just mouth movement. Eyes, cheeks, and expressions move with the audio. Eyebrows raise on emphasis, the jaw opens and closes naturally, and the head tilts to match the intonation of the speech. The whole face participates in the animation, not just the mouth.
Script the voiceover, have talent record it, upload a photo, and Craze generates the talking video. No green screen, no camera crew, no sync work. This workflow is much faster than traditional video shooting, and it works with anyone's image.
No upscaling, no separate audio sync. LTX-2.5 understands 1080p talking heads from the ground up. Lightricks built this model specifically for the talking photo use case, and it shows in the quality.
LTX-2.5 specializes in audio-to-video animation. You provide a portrait photo and an audio file, and the model animates the face to match the speech. The model understands phonetics, facial mechanics, and natural animation. It generates tightly synced lip movement and realistic facial expressions that match the emotional tone of the audio.
LTX-2.5 works with still images: headshots, portraits, close-ups of faces, or character artwork. The image should be front-facing and show the face clearly. The audio can be a voiceover, dialogue, narration, or any speech recording. The model will animate the image to sync with that audio.
Lightricks released LTX-2.5 as an open-weight model in August 2026. On Craze, you just upload your image and audio, name LTX-2.5, and Craze handles the generation.
You pay per second of audio. $0.179 per second of audio. Longer audio means longer video and higher cost. The cost calculator on this page shows the exact price for your clip.
Audio-to-video animation scales with the audio duration, so a 10-second voiceover costs more than a shorter one. The pricing is straightforward and doesn't vary by image quality or other parameters.
Record a script, upload a headshot, and LTX-2.5 animates it. No on-camera talent needed. Works for product narration, brand messaging, or any voiceover-driven content.
Convert interview audio into a video. Use a still image of the person and LTX-2.5 makes it look like they're speaking on camera. For case studies, testimonials, or interview content, you just need the audio and a photo.
A subject matter expert records narration. LTX-2.5 animates their image to match, and you have a professional-looking tutorial without video equipment. For online courses, training videos, and educational content, this is cost-effective.
Build a presentation where an animated avatar delivers the speech. Use a real photo or a generated character image, and LTX-2.5 animates it to your audio. For virtual presentations, webinars, and conference talks, this creates a consistent visual anchor.
When you can't get talent on set, have them record audio and use their photo. LTX-2.5 handles the rest. This is valuable for distributed teams and remote collaboration.
LTX-2.5 is specialized for audio-to-video talking photos. It's built specifically for this task and handles lip sync and facial animation well. For full video generation with scenes, environments, and characters in motion, see LTX-2.3 for 4K multishot work, Gemini Omni Flash for multimodal flexibility, or Veo 3.1 for maximum resolution and video extension.
Describe what you want the way you would brief an editor, and Craze plans the shots and writes the LTX 2.5 prompts for you. Every Craze plan, including Free, runs LTX 2.5 and every other media model, and Craze can use a different one for each job. Each video lands on an editable timeline with your cuts, captions and music, ready to export as MP4. Generations come out of your plan's monthly credits, from $0.179 per second of audio, with top-ups when you need more.
Voiceover and narration videos: Record a script, upload a headshot, and LTX-2.5 animates it. Interviews and testimony: Convert interview audio into a video. Educational and tutorial content: A subject matter expert records narration. Accessible avatars for presentations: Build a presentation where an animated avatar delivers the speech.
LTX-2.5 is Lightricks' talking photo model. It takes a still image and audio, and animates the image to create a talking head video at 1080p. Each clip syncs mouth movements to the audio.
Lightricks released LTX-2.5 on August 11, 2026, as an open-weight model with day-one support in multiple platforms including ComfyUI.
1080p. LTX-2.5 is purpose-built for talking photos at full HD resolution.
LTX-2.5 creates talking head videos from a still image and audio. You provide a photo and a voiceover or speech recording, and it animates the face to match the audio.
Yes. You provide the audio separately as a voiceover, dialogue recording, or other speech. LTX-2.5 then syncs the face animation to that audio.
LTX-2.5 is trained specifically for accurate mouth and facial movement. Every phoneme is animated to match the audio, so the lip sync is tight.
LTX-2.5 works with portrait images, headshots, and head-only crops. Give it a clear front-facing photo of the person or character you want to animate.
$0.179 per second of audio. The rate depends on the duration of the audio clip. The cost calculator on this page shows exact pricing.
every Craze plan, including Free. Every paid plan includes monthly credits that cover LTX-2.5 generations.
ByteDance Seed Team
From $0.114 per second of video
ByteDance Seed Team
From $0.075 per second of video
ByteDance Seed Team
From $0.06 per second of video
