> For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://docs.empiriolabs.ai/models/pixverse-v6/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.empiriolabs.ai/_mcp/server. # Pixverse v6 > Generates video from text, images, first and last frames, or up to seven references, with native audio, multi-shot cuts, and clip extension. ![Pixverse v6](https://media.empiriolabs.ai/model-logos/pixverse-v6.png) [PixVerse](/providers/pixverse) · Video Generation `POST /v1/videos/generations` Generates video from text, images, first and last frames, or up to seven references, with native audio, multi-shot cuts, and clip extension. ## At a glance | Field | Value | | ------------------- | --------------------------------------------------------------------------------------------------------------------- | | Model id | `pixverse-v6` | | Model release date | 2026-03-29 | | Input modalities | Text, Image, Video | | Output modalities | Video | | Context window | - | | Weight precision | - | | Features | text\_to\_video, image\_to\_video, reference\_to\_video, audio\_sync, multi\_shot, character\_consistency, multimodal | | Native inference | No | | New | No | | Supported endpoints | `POST /v1/videos/generations` | | Alternate model ids | `pixverse/v6` | ## Pricing | Charge | Spec | Rate | | -------------------------------- | ---------- | ---------- | | Standard 360p no audio | per second | \$0.053333 | | Standard 360p with audio | per second | \$0.074667 | | Standard 540p no audio | per second | \$0.074667 | | Standard 540p with audio | per second | \$0.096 | | Standard 720p no audio | per second | \$0.096 | | Standard 720p with audio | per second | \$0.128 | | Standard 1080p no audio | per second | \$0.192 | | Standard 1080p with audio | per second | \$0.245333 | | Video reference 360p no audio | per second | \$0.106667 | | Video reference 360p with audio | per second | \$0.149333 | | Video reference 540p no audio | per second | \$0.149333 | | Video reference 540p with audio | per second | \$0.192 | | Video reference 720p no audio | per second | \$0.192 | | Video reference 720p with audio | per second | \$0.256 | | Video reference 1080p no audio | per second | \$0.384 | | Video reference 1080p with audio | per second | \$0.490667 | ## Example request ```bash curl https://api.empiriolabs.ai/v1/videos/generations \ -H 'Authorization: Bearer $EMPIRIOLABS_API_KEY' \ -H 'Content-Type: application/json' \ -d '{"model": "pixverse-v6", "prompt": "sunrise over the ocean", "duration": 6}' ``` ## Parameters | Parameter | Type | Required | Default | Description | | ---------------------------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | `prompt` | string | yes | - | Scene description, up to 2048 characters. In reference mode you can address a named reference inline as @name followed by a space. | | `mode` | enum | no | `"auto"` | How your inputs are used. auto detects it from what you attach. t2v is text to video. i2v animates one image. transition interpolates between a first and last frame. reference combines up to 7 reference images, optionally with reference videos. extend continues an uploaded clip. · Allowed: `auto`, `t2v`, `i2v`, `transition`, `reference`, `extend` | | `resolution` | enum | no | `"720p"` | Output resolution. Higher resolutions bill at a higher per-second rate. · Allowed: `360p`, `540p`, `720p`, `1080p` | | `aspect_ratio` | enum | no | `"16:9"` | Output aspect ratio. Used by text-to-video and reference modes. Image-driven modes follow the source image instead. · Allowed: `16:9`, `4:3`, `1:1`, `3:4`, `9:16`, `2:3`, `3:2`, `21:9` | | `duration` | number | no | `5` | Clip length in seconds. Ignored when a reference video is supplied, because the output length follows the longest reference. · Range: 1 – 15 | | `generate_audio_switch` | boolean | no | false | Generate native synchronized audio with the video. Bills at the higher with-audio rate for the chosen resolution. | | `generate_multi_clip_switch` | boolean | no | false | Produce a multi-shot cut instead of one continuous take. Text-to-video and image-to-video only. | | `seed` | number | no | - | Reproducibility seed. The same seed with the same inputs returns a similar result. · Range: 0 – 2147483647 | | `reference_names` | string | no | - | Comma-separated names for the attached reference images, in order, so the prompt can address them as @name. | | `video_reference_names` | string | no | - | Comma-separated names for the attached reference videos, in order. Supplying reference videos switches the model into omni reference and bills at the video-reference rate. | | `image` | string | no | - | Source or first-frame image URL. | | `image_end` | string | no | - | Last-frame image URL for transition mode. | | `video` | string | no | - | Source video URL for extend, or a reference video URL for reference mode. | ## Notes PixVerse V6 generates video from a text prompt, a single image, a first and last frame, or up to seven reference images, and can continue an existing clip. **Modes** * `t2v`: text to video. * `i2v`: animate a single image. * `transition`: interpolate between a first-frame and a last-frame image. * `reference`: combine up to 7 reference images, optionally with up to 3 reference videos. * `extend`: continue an uploaded video. **Output** * 360p, 540p, 720p and 1080p, 1 to 15 seconds. * Aspect ratios 16:9, 4:3, 1:1, 3:4, 9:16, 2:3, 3:2 and 21:9 on text-to-video and reference modes. Image-driven modes follow the source image. * Set `generate_audio_switch` to true for native synchronized audio. * Set `generate_multi_clip_switch` to true on text-to-video and image-to-video for a multi-shot cut. **Reference videos** * Supplying `videos` in `reference` mode switches the model into omni reference. Output length is taken from the longest reference video, so `duration` is ignored. * Name a reference with `reference_names` or `video_reference_names` and address it in the prompt as `@name` followed by a space. **Billing** * Billed per second of generated video at the rate for the output resolution, with a higher rate when audio is on. * A reference request that carries a reference video bills at the video-reference rate, which is roughly double the standard rate. **Input limits** * Images: PNG, JPEG or WebP, up to 20MB and 10000px on the long edge. * Videos for extend: MP4, MOV or WebM, up to 30 seconds and 50MB. --- *Machine-readable schema:* `GET https://api.empiriolabs.ai/v1/models/pixverse-v6`. > Generates video from text, images, first and last frames, or up to seven references, with native audio, multi-shot cuts, and clip extension.