flux-3-maxFLUX 3 Max is a post-trained, latency-optimized variant of FLUX 3 Video, the video model family from Black Forest Labs, tuned for stronger prompt adherence and better visual quality. It is built to return clips fast: in our own test on 2026-10-20 a 5-second 720p text-to-video clip was ready 9 seconds after submission through the API, delivered as a 1280x704, 24 fps file with a synced audio track. The vendor's own benchmark puts a 10-second video at about 11 seconds, and according to the vendor the model placed joint first in a preliminary Artificial Analysis image-to-video evaluation (rankings can change). One model id, flux-3-max, covers three modes, chosen from what you send. A prompt alone is text-to-video. A first frame (first_frame_url, also accepted as image or image_url) makes it image-to-video, and an optional last_frame_url sets where the clip ends. Keyframes go further: pass up to 10 images in keyframes, each pinned to a moment by frame_index (a frame number at 24 fps) or by time in seconds, and the model builds the motion between them; keyframes cannot be combined with a first or last frame. Every mode takes duration as any whole number of seconds from 5 to 20 (default 5), resolution 720p or 1080p (default 720p), and aspect_ratio auto, 21:9, 2:1, 16:9, 4:3, 1:1, 3:4 or 9:16, where auto means 16:9 for text-to-video and follows the first image otherwise. Audio is generated by default and can be turned off with generate_audio: false; reasoning_level is high by default, and low is faster; seed is optional. Reference videos, reference audio and more than one reference image outside keyframes are not accepted and return a clear 400 instead of being silently dropped. Billing is per second of output: 720p at $0.221 per second and 1080p at $0.377 per second, so a 5-second 720p clip is $1.105 and a 10-second 1080p clip is $3.77, with a special price of $0.085/s (720p) and $0.145/s (1080p) until November 7, 2026. Only successful generations are charged; failed ones are refunded.
Auto = 16:9 for text-to-video
Synced sound is generated with the clip
Default
720p $0.221/s · 1080p $0.377/s · 5-20s · only successful generations are billed
Generated video will appear here
Provide URLs and click Generate
Our own test on 2026-10-20: a 5-second 720p text-to-video clip was ready 9 seconds after submission through the API, at 1280x704 and 24 fps with audio. The vendor benchmarks a 10-second video at about 11 seconds
A prompt alone is text-to-video; first_frame_url plus an optional last_frame_url is image-to-video; keyframes is keyframes-to-video. The mode follows the inputs you send
Pin each image to a moment with frame_index (24 fps) or time in seconds and the model fills in the motion between them. Keyframes cannot be mixed with a first or last frame
Every clip comes with a generated soundtrack unless you send generate_audio: false
Any whole length from 5 to 20 seconds, aspect ratios from 21:9 to 9:16 or auto, and reasoning_level low when speed matters most
720p $0.221/s and 1080p $0.377/s for the seconds you request, with a special price of $0.085/s and $0.145/s until Nov 7. Failed generations are refunded
FLUX 3 Max is a Video Generation API provided by Black Forest Labs. FLUX 3 Max is a post-trained, latency-optimized variant of FLUX 3 Video, the video model family from Black Forest Labs, tuned for stronger prompt adherence and better visual quality. It is built to return clips fast: in our own test on 2026-10-20 a 5-second 720p text-to-video clip was ready 9 seconds after submission through the API, delivered as a 1280x704, 24 fps file with a synced audio track. The vendor's own benchmark puts a 10-second video at about 11 seconds, and according to the vendor the model placed joint first in a preliminary Artificial Analysis image-to-video evaluation (rankings can change). One model id, flux-3-max, covers three modes, chosen from what you send. A prompt alone is text-to-video. A first frame (first_frame_url, also accepted as image or image_url) makes it image-to-video, and an optional last_frame_url sets where the clip ends. Keyframes go further: pass up to 10 images in keyframes, each pinned to a moment by frame_index (a frame number at 24 fps) or by time in seconds, and the model builds the motion between them; keyframes cannot be combined with a first or last frame. Every mode takes duration as any whole number of seconds from 5 to 20 (default 5), resolution 720p or 1080p (default 720p), and aspect_ratio auto, 21:9, 2:1, 16:9, 4:3, 1:1, 3:4 or 9:16, where auto means 16:9 for text-to-video and follows the first image otherwise. Audio is generated by default and can be turned off with generate_audio: false; reasoning_level is high by default, and low is faster; seed is optional. Reference videos, reference audio and more than one reference image outside keyframes are not accepted and return a clear 400 instead of being silently dropped. Billing is per second of output: 720p at $0.221 per second and 1080p at $0.377 per second, so a 5-second 720p clip is $1.105 and a 10-second 1080p clip is $3.77, with a special price of $0.085/s (720p) and $0.145/s (1080p) until November 7, 2026. Only successful generations are charged; failed ones are refunded. Through APIMODELS platform, you can access this model via a unified API with transparent pay-as-you-go pricing. Current pricing: 720p · per second: $0.221, 720p · 5s: $1.105, 720p · 10s: $2.21, 720p · 20s: $4.42, 1080p · per second: $0.377, 1080p · 5s: $1.885, 1080p · 10s: $3.77, 1080p · 20s: $7.54.
Related model: minimax-h3-max-turbo — the other seconds-fast video tier — up to 768P and 15 seconds at $0.096/s, text-to-video or first/last frame
Related model: gemini-omni-1.1-flash — when you need 4K, up to 7 reference images, or an edit of an existing clip
Related model: minimax-h3-rh-enhanced — when the shot needs reference videos, or your own audio track to drive lip sync
Three clips generated on this exact model on 2026-10-20, the day it went live, through the same public endpoint you would call: one per mode — text-to-video, image-to-video from a product photo, and keyframes-to-video between two images. Times are wall-clock from the create request to a downloadable file (9, 15 and 19 seconds), the prompt under each clip is exactly what produced it, and the tier label is the regular price with the special price in brackets. Nothing was cherry-picked; these are the first three runs. For parameters, pricing and a comparison with Seedance 2.5, see the FLUX 3 Max API guide.
A red paper boat drifts down a rain-soaked city gutter at dusk, neon reflections ripple on the water, camera follows at street level, soft rain sounds
Delivered as 1280×704 at 24 fps with an audio track; aspect_ratio 16:9, reasoning_level low.
The mug slowly turns on the table as thin steam rises, soft morning light, gentle camera push in, quiet kitchen ambience
aspect_ratio auto followed the square first frame, so the file is 960×960, with audio.
Smooth cinematic transition from a casual kitchen snapshot of the mug to a premium studio product shot on dark stone, light sweeps across the mug
keyframes: [{time: 0}, {time: 5}] (frames 0 and 120 at 24 fps), generate_audio false, so the file has no audio track.
Quickly generate brand promotion videos for ad campaigns and social media marketing.
Create compelling short-form video content for platforms like TikTok, Instagram, and YouTube.
Generate product feature demonstrations and tutorials to improve user conversion.
Produce course explanations, knowledge explainers, and training videos at an affordable price.
FLUX 3 Max is available through APIMODELS at: 720p · per second: $0.221, 720p · 5s: $1.105, 720p · 10s: $2.21, 720p · 20s: $4.42, 1080p · per second: $0.377, 1080p · 5s: $1.885, 1080p · 10s: $3.77, 1080p · 20s: $7.54. Billing is pay-as-you-go — you only pay for what you generate.
Sign up at APIMODELS, get your API key, and call our unified API endpoint. We provide detailed API documentation with code examples in cURL, Python, and Node.js.
APIMODELS offers the same FLUX 3 Max model through our aggregation platform. We provide a unified API interface so you do not need separate accounts for each provider - one API key to access all models.
FLUX 3 Max is a post-trained, latency-optimized variant of FLUX 3 Video, the video model family from Black Forest Labs, tuned for stronger prompt adherence and better visual quality. On apimodels.app it is one model id, flux-3-max, on the unified video endpoint, covering text-to-video, first/last-frame image-to-video and keyframes-to-video at 720p or 1080p, 5 to 20 seconds, with synced audio generated by default.
Per second of output: 720p at $0.221/s (5 s $1.105, 10 s $2.21, 20 s $4.42) and 1080p at $0.377/s (5 s $1.885, 10 s $3.77, 20 s $7.54). Until November 7, 2026 a special price of $0.085/s (720p) and $0.145/s (1080p) applies, and the regular price returns automatically after that. Only successful generations are charged; failed ones are refunded.
POST https://api.apimodels.app/v1/video/generations with model flux-3-max and a prompt. Optional: duration, 5-20 whole seconds (default 5); resolution, 720p or 1080p (default 720p); aspect_ratio, one of auto, 21:9, 2:1, 16:9, 4:3, 1:1, 3:4, 9:16 (default auto); generate_audio (default true); reasoning_level, high or low (default high); seed. Add first_frame_url (and optionally last_frame_url) for image-to-video, or keyframes for keyframes-to-video. The call returns a taskId; poll GET /v1/video/generations?task_id= until it is completed, or pass callback_url.
Send keyframes as an array of 1 to 10 objects, each with an image_url and either frame_index (the frame number in a 24 fps video, from 0 to duration × 24) or time in seconds, which we convert to a frame index. No two keyframes may land on the same frame. The model treats each image as a point the video has to pass through and generates the motion between them. Keyframes replace the first/last frame inputs: sending both returns a 400.
In our own test on 2026-10-20, a 5-second 720p text-to-video clip was ready 9 seconds after submission through the API, as a 1280x704, 24 fps file with audio. The vendor's benchmark puts a 10-second video at about 11 seconds, and the vendor reports joint first place in a preliminary Artificial Analysis image-to-video evaluation; rankings can change. 1080p, longer clips and requests with image inputs were not part of our test.
Reference videos, reference audio, and more than one reference image outside keyframes. A request carrying any of those gets a clear 400 at create time instead of being silently dropped and billed. One image goes in first_frame_url (image, image_url or a single-item images array are accepted as the first frame too); several images go in keyframes.
On APIMODELS, FLUX 3 Max runs alongside 60+ models on one API key and one balance, so choosing is about fit, not lock-in. It supports Clips in Seconds, Text-to-Video, First / Last Frame, Up to 10 Keyframes, Synced Audio, 720p / 1080p, 5-20s, From $0.221/s, and you can weigh it on price and capability against other Video Generation models, then switch by changing a single model-name string — no new account or integration. Browse every Video Generation option with live pricing at apimodels.app/models.
FLUX 3 Max supports: Clips in Seconds, Text-to-Video, First / Last Frame, Up to 10 Keyframes, Synced Audio, 720p / 1080p, 5-20s, From $0.221/s. See the APIMODELS docs for full parameters and call examples.
Yes. APIMODELS exposes FLUX 3 Max through a single unified API and one key — no separate provider accounts, and no need to handle each provider's regional network access yourself.
We support Stripe (Visa, Mastercard, and other international cards) and Alipay. Credits are available instantly after payment.
How to get access, regional availability, and how this model compares with its alternatives.