FLUX 3 Max AI video API: budget at the post-promo price
fal launched FLUX 3 Max, an AI video API at $0.17 per second at 720p, 75% off until November 7. Price product video at the full rate before you scale.
fal and Black Forest Labs have put a new AI video model on a public API. FLUX 3 Max makes 5 to 20 second clips from text, an image, or up to 10 keyframes. fal claims it renders a 10-second clip in about 12 seconds. For a store that needs product video, the headline is the price: about 43 cents per 10-second 720p clip. That price has an end date.
What actually happened
Per VentureBeat and fal's model page:
- Three endpoints: text-to-video, image-to-video (with an optional end frame), and keyframes-to-video.
- The output: 720p native with optional upscaling to 1080p, audio on or off, and landscape, portrait, or square.
- The price: $0.0425 per second at 720p and $0.0725 at 1080p. That is a 75% launch discount that runs until November 7. After that, it is $0.17 per second at 720p and $0.29 at 1080p.
- The speed claim: 12.45 seconds to make a 10-second clip, against 277.80 seconds for Seedance 2.5. That number is fal's own benchmark, not an independent test.
- The quality: On the Artificial Analysis image-to-video leaderboard, FLUX 3 Max sits in a near-tie with MiniMax H3 and Seedance 2.5. The confidence ranges overlap, so it is not a clear winner.
This is a step on from the gated FLUX 3 Video preview we covered earlier. You can now call it from your own code.
Why AI video API pricing matters for your business
Budget at the regular rate. A 10-second 720p clip costs $0.425 now and $1.70 after November 7. You will not get a usable shot on the first try. Ten tries at full price is $17 per shot. For 200 product pages, plan for the full rate, not the launch rate.
Speed changes the workflow. If a clip takes 12 seconds instead of 5 minutes, someone can review and re-prompt in one sitting. That makes a batch pipeline possible: product photo in, three candidate clips out, a human picks one.
Keep the model swappable. Three models are in a statistical tie on quality today. Next month the ranking will change, and so will the prices. Put the model call behind one function in your pipeline, so changing vendors is a config edit.
Read the usage terms. fal says API output can be used commercially, subject to its terms. Check them before the clips go into paid ads.
Key takeaways
- FLUX 3 Max is live on fal's API for text-, image-, and keyframe-to-video, 5 to 20 seconds
- 720p costs $0.0425/second until November 7, then $0.17/second; 1080p goes from $0.0725 to $0.29
- The 22x speed claim is fal's own benchmark; quality is in a near-tie with two rivals
- Budget at the full rate and assume several attempts per usable shot
- Put the video model behind one swappable function in your pipeline
Need video for a full catalog, not one hero clip? We build product-video pipelines that take your existing photos, call whichever model wins on price that month, and route clips to a human for approval. See what we build, or price it against your current production cost.
Sources: VentureBeat, fal.
- #ai-video
- #flux-3-max
- #fal
- #product-video
- #api-pricing
Tommy Rush — Founder, Rush Commerce
Operator turned builder. 15+ years running operations — now shipping the systems businesses run on. More
Get The Rush Report weekly — one email, zero fluff.
Keep reading
Mercury 3 hits 1,068 tokens/sec. Don't price on the discount.
Inception's Mercury 3 diffusion LLM runs 1,068 tokens/sec with a 1M context at $0.20/$0.70. The 50% launch discount has no end date — budget at list.
Read itdoubleAI Maxwell rewrites running code. Profile before you pay.
doubleAI's Maxwell rewrites machine code under running Python and charges a share of compute savings. Profile your workload before you sign.
Read it