FLUX 3 Max on Segmind: text and image to video with native audio
FLUX 3 Max is a post-trained version of Black Forest Labs' FLUX 3. Generate up to 20-second videos with native synchronized audio from a text prompt, or animate a still first frame into a moving clip — both at 720p or 1080p.
FLUX 3 Max is a post-trained version of Black Forest Labs' FLUX 3 multimodal model, and it reaches Segmind as two modes. Text to video builds a clip from a written prompt, while image to video animates a still first frame you supply. Both return up to 20 seconds at 720p or 1080p with native synchronized audio in the same pass, so a sound lands on the frame of its event. The model is built for speed — a 10-second clip returns in roughly 12 seconds — and both modes share duration, aspect-ratio, reasoning-level and seed controls.
Text to video
FLUX 3 Max Text to Video turns a written description of subject, action, camera move and sound into a finished clip, with no starting image. Lead with a concrete subject, give it one clear action, then a camera move that the action motivates. Finish with one line of sound: because audio is generated in the same pass, naming the sound ties it to the on-screen event. Specific nouns and a single, legible motion read better than stacked adjectives.
It holds motion together on busy, multi-subject scenes, and renders faces and materials that stay stable in close-up. Keep reasoning level high for complex prompts and drop to low for simple scenes, and start at 5 seconds while you iterate before extending the duration.
ModelFLUX 3 Max Text to VideoText-to-video with native audio, up to 20 seconds, 1080p.Each clip below ran end to end at the settings on its card, so the motion, framing and audio all came from the prompt. Click a card to drop its prompt and settings back into the form.
Three taiko drummers in indigo happi coats drive their wooden bachi into a huge nagado-daiko and two shime-daiko in unison, lunging low on every downbeat, sweat flicking off their arms; a hanging red paper lantern sways behind them on a wooden matsuri stage. The camera dollies laterally past them at the height of the drum skins. Audio: thunderous synchronized drum booms locked to each strike, sharp rim clacks, a shouted 'Sore!', a crowd murmuring in the dark.
Three taiko drummers strike in unison on a night matsuri stage and each drum boom lands on the frame it is struck, with the sticks holding their shape. Tip: name the sound you want in an 'Audio:' line and keep generate_audio on.
Four thoroughbred racehorses thunder around the final dirt bend of a racetrack, jockeys crouched low in bright silks, clods of earth flung up behind the hooves; the camera tracks alongside at rail height, level with the leader's shoulder. Audio: pounding hooves in a quickening rhythm, a jockey's sharp shout, a rising crowd roar.
Four horses gallop round the bend with legs, dust and jockey silks reading cleanly at speed and hoofbeats locked to the stride. Tip: keep reasoning_level high for busy multi-subject motion and track the camera with the action; start at 5 s before extending.
Extreme close-up of an elderly watchmaker's face and hands, a brass jeweler's loupe screwed into one eye, as she lowers hair-thin tweezers to seat a tiny gold gear into an open pocket-watch movement; daylight from a workshop window rakes across her lined skin and the polished brass. Audio: a faint mechanical ticking, the tiny click of the tweezers, a quiet room tone.
An extreme close-up holds natural, lined skin and a jeweled gold pocket-watch movement at close range while facial features stay stable. Tip: close-ups reward a single quiet action and one line of incidental sound; switch to 1080p for final delivery.
Use text to video from the API
Send a POST with your prompt and settings, and the response returns the finished clip. This is the default example:
# pip install "segmind>=1.1.0"
# export SEGMIND_API_KEY="YOUR_API_KEY"
import segmind
result = segmind.run(
"flux-3-max-text-to-video",
prompt="Three taiko drummers in indigo happi coats drive their wooden bachi into a huge nagado-daiko and two shime-daiko in unison, lunging low on every downbeat, sweat flicking off their arms; a hanging red paper lantern sways behind them on a wooden matsuri stage. The camera dollies laterally past them at the height of the drum skins. Audio: thunderous synchronized drum booms locked to each strike, sharp rim clacks, a shouted 'Sore!', a crowd murmuring in the dark.",
duration=5,
resolution="720p",
aspect_ratio="16:9",
generate_audio=True,
reasoning_level="high",
seed=94355960,
)
print(result.get("output"))# 1. Submit the request; the reply carries its request_id
REQUEST_ID=$(curl -s -X POST "https://api.segmind.com/v2/flux-3-max-text-to-video" \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"prompt": "Three taiko drummers in indigo happi coats drive their wooden bachi into a huge nagado-daiko and two shime-daiko in unison, lunging low on every downbeat, sweat flicking off their arms; a hanging red paper lantern sways behind them on a wooden matsuri stage. The camera dollies laterally past them at the height of the drum skins. Audio: thunderous synchronized drum booms locked to each strike, sharp rim clacks, a shouted '\''Sore!'\'', a crowd murmuring in the dark.",
"duration": 5,
"resolution": "720p",
"aspect_ratio": "16:9",
"generate_audio": true,
"reasoning_level": "high",
"seed": 94355960
}' | jq -r .request_id)
# 2. Check its status until it is COMPLETED
curl -s "https://api.segmind.com/v2/requests/$REQUEST_ID/status" \
-H "x-api-key: YOUR_API_KEY"
# 3. Fetch the result
curl -s "https://api.segmind.com/v2/requests/$REQUEST_ID" \
-H "x-api-key: YOUR_API_KEY"Image to video
FLUX 3 Max Image to Video animates a still you supply — a PNG, JPEG or WebP first frame — into up to 20 seconds of video with native synchronized audio. The clip begins exactly on that frame, so compose it at the top of an action to give the motion somewhere to go. In the prompt, describe what moves, a camera move that the action motivates, and one line of sound. Fast subjects keep the shape they had in the first frame instead of smearing or morphing.
Supply a last frame as well and the model interpolates a controlled first-to-last transition. A clean, well-lit frame with a clear subject animates more predictably than a cluttered one.
ModelFLUX 3 Max Image to VideoAnimate a photo into video with native synchronized audio.Each clip below animates the first frame shown at the settings on its card, with the motion, camera move and sound coming from the prompt.
The three drummers drive their raised bachi down into the drums in unison, lunging low on the downbeat, arms snapping through the strike and sweat flicking off them; the red paper lantern sways behind them. A low lateral dolly tracks past the drum skins at the moment of impact. Audio: thunderous synchronized taiko booms locked to each strike, sharp rim clacks, a shouted 'Sore!', a crowd murmuring in the dark.
Three taiko drummers drive the downbeat in unison; arms and bachi hold their shape through the motion blur and the drum booms land on each visible strike. Tip: compose the first frame at the top of the action so the motion has somewhere to go.
The four racehorses surge around the dirt bend in a tight pack, legs scissoring and clods of earth flung up behind the hooves, the jockeys rising and dropping in their bright silks; the camera tracks alongside at rail height, level with the leader's shoulder. Audio: a drumming gallop of hooves on packed dirt, short shouted calls from the jockeys, a distant grandstand roar.
The four horses gallop and the camera tracks past at rail height while legs, jockeys and flung dirt keep their shape through the motion blur. Tip: shoot the first still at full stride and keep reasoning_level high for multi-subject motion.
The watchmaker lowers the hair-thin tweezers and seats the tiny gold gear into the open movement, his loupe eye narrowing as he holds his breath; a slow push-in stays on his hands and the brass wheels. Audio: a faint metallic click as the gear beds in, the soft tick of the movement starting, quiet room tone.
He lowers the tweezers to seat the gold gear and the push-in holds on his hands while skin, the brass loupe and the movement's wheels stay sharp. Tip: start from a well-lit close-up and move to 1080p for a final.
Use image to video from the API
Send a POST with your first frame, prompt and settings, and the response returns the finished clip. This is the default example:
# pip install "segmind>=1.1.0"
# export SEGMIND_API_KEY="YOUR_API_KEY"
import segmind
result = segmind.run(
"flux-3-max-image-to-video",
image="https://segmind-resources.s3.amazonaws.com/input/flux-3-max-image-to-video-firstframe.jpg",
prompt="The three drummers drive their raised bachi down into the drums in unison, lunging low on the downbeat, arms snapping through the strike and sweat flicking off them; the red paper lantern sways behind them. A low lateral dolly tracks past the drum skins at the moment of impact. Audio: thunderous synchronized taiko booms locked to each strike, sharp rim clacks, a shouted 'Sore!', a crowd murmuring in the dark.",
duration=5,
resolution="720p",
aspect_ratio="16:9",
generate_audio=True,
reasoning_level="high",
seed=824610357,
)
print(result.get("output"))# 1. Submit the request; the reply carries its request_id
REQUEST_ID=$(curl -s -X POST "https://api.segmind.com/v2/flux-3-max-image-to-video" \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"image": "https://segmind-resources.s3.amazonaws.com/input/flux-3-max-image-to-video-firstframe.jpg",
"prompt": "The three drummers drive their raised bachi down into the drums in unison, lunging low on the downbeat, arms snapping through the strike and sweat flicking off them; the red paper lantern sways behind them. A low lateral dolly tracks past the drum skins at the moment of impact. Audio: thunderous synchronized taiko booms locked to each strike, sharp rim clacks, a shouted '\''Sore!'\'', a crowd murmuring in the dark.",
"duration": 5,
"resolution": "720p",
"aspect_ratio": "16:9",
"generate_audio": true,
"reasoning_level": "high",
"seed": 824610357
}' | jq -r .request_id)
# 2. Check its status until it is COMPLETED
curl -s "https://api.segmind.com/v2/requests/$REQUEST_ID/status" \
-H "x-api-key: YOUR_API_KEY"
# 3. Fetch the result
curl -s "https://api.segmind.com/v2/requests/$REQUEST_ID" \
-H "x-api-key: YOUR_API_KEY"Shared settings
These controls are shared across both modes; a few inputs belong to image to video only.
| Setting | Value | Members | When |
|---|---|---|---|
| prompt | Subject or motion, one action, a camera move, one line of sound | Both | Required; it drives the motion, framing and audio |
| image | PNG, JPEG or WebP first frame | Image to video | Required; compose it at the top of an action |
| last_frame_image | PNG, JPEG or WebP | Image to video | Optional; set it for a controlled first-to-last transition |
| duration | 5–20 s (default 5) | Both | Start at 5 s to iterate, raise it for the final |
| resolution | 720p (default) or 1080p | Both | 720p for drafts; 1080p for final delivery |
| aspect_ratio | auto, 21:9, 2:1, 16:9, 4:3, 1:1, 3:4, 9:16 | Both | auto is 16:9 for text and follows the image for image to video; use 9:16 for social |
| generate_audio | true (default) | Both | Keep on for any sound event; off for a silent clip |
| reasoning_level | low or high (default high) | Both | high for busy multi-subject prompts; low is faster for simple scenes |
| seed | -1 random, or a fixed integer | Both | Fix it to reproduce an exact clip |
Good to know
- Output is 720p or 1080p. Some base FLUX 3 Video features — 4K, keyframe-to-video and video-to-video — are not on these two endpoints.
- On Artificial Analysis's preliminary image-to-video arena, FLUX 3 Max is tied for the top score within the confidence interval. The arena is company-run and preliminary, so read it as frontier quality, not a settled ranking.
- Image to video begins exactly on your first frame, so a clean, well-lit frame with a clear subject animates more predictably than a cluttered one.
- Audio is on by default and rendered with the picture. Name the sound you want in an "Audio:" line, and switch generate_audio off for a silent clip.
Resources
- Try FLUX 3 Max Text to Video on Segmind · API reference
- Try FLUX 3 Max Image to Video on Segmind · API reference
- FLUX 3 Video model card (Black Forest Labs) — the base FLUX 3 Video model card: multimodal architecture, text, image, video and keyframe inputs, resolution bands up to 4K, and native audio
- FLUX 3: Multimodal Video, Image & Audio (Black Forest Labs) — the technical announcement of FLUX 3 and the Self-Flow approach: joint image, video and audio training, native synchronized audio and multilingual speech
- #flux-3-max
- #black-forest-labs
- #model-launch
- #video-generation
- #native-audio
