These three controls decide how good your result looks, how big it is, and how long it runs. They also affect cost — and in many modes they constrain each other, so changing one can limit (or auto-adjust) another.
What does the Quality control do?
Quality picks the model tier used for your generation. Higher tiers generally look better and cost more. The tiers you'll see depend on the mode — image lanes use Fast, Ultra 2K, Ultra 4K, and UltraS 4K, while video lanes use Fast, Ultra, and UltraS.
Each chip has a hover tooltip with a short description and the per-tier cost.
Which quality tiers does each mode offer?
It varies by mode:
Generate Image — Fast, Ultra 2K, Ultra 4K, UltraS 4K (default Ultra 4K).
Edit Image — only Chat to edit exposes a quality choice (Ultra 4K, UltraS 4K); Face swap, Edit clothing, Upscale, and Remove background have no quality picker and run at a fixed quality.
Image to video — Fast, Ultra, UltraS (paid-user default UltraS).
Text to video — Fast, Ultra, UltraS (paid-user default UltraS).
Extend video — Fast, Ultra, UltraS (paid-user default UltraS).
Storyboard to video — Ultra, UltraS.
Clone Video — Ultra, UltraS (default UltraS).
Edit Video — the exact set depends on the editing tool; paid users start in Edit character on UltraS.
The exact default and offered set can depend on your account and plan, so your list may differ slightly.
Why are some quality tiers missing for me?
The available tiers differ per mode (see above) — the image lanes expose four tiers, the standard video lanes three, and the storyboard lane only two. Your account and plan can also affect which tier is selected by default.
What does Resolution do, and how is it tied to quality?
Resolution sets the output size for video modes. Options use the labels 480P (SD), 720P (HD), FHD, QHD, and 4K, and which ones a lane offers varies. Resolution is also constrained by your quality tier: a resolution that isn't available at your current quality is shown disabled with a tooltip like "Only available for …". The standard video lanes work like this:
Image to Video — paid users start on UltraS 720P. 480P is available only on Fast and is the free-user default; that MiniMax H3 option supports 4-second, 5-second, and 10-second output. On desktop, the menu still shows unavailable 15s, 20s, 25s, and 30s options as disabled with the quality tiers that unlock them; on mobile, unavailable durations are filtered out, so Fast shows 4s, 5s, and 10s. 720P, FHD, and QHD are available on every tier (Fast, Ultra, UltraS), while 4K is available on Ultra and UltraS only. Fast's quality tooltip identifies "Powered by MiniMax H3".
Text to Video — paid users start on UltraS 720P. 480P is available only on Fast and is the free-user default; Fast supports 4-second, 5-second, and 10-second output. 720P, FHD, and QHD are available on every tier (Fast, Ultra, UltraS), while 4K is available on Ultra and UltraS only. Fast's quality tooltip identifies "Powered by MiniMax H3".
Extend Video — paid users start on UltraS 720P, while free users retain the established Fast 720P starting selection. 480P is available only on Fast, which also supports 720P and FHD. Fast supports 4-second, 5-second, and 10-second output, and its quality tooltip identifies "Powered by MiniMax H3". Ultra and UltraS continue to offer 720P and FHD.
Storyboard to video — 720P, FHD, and QHD work on either tier (Ultra, UltraS); 4K is available only at the highest tier, UltraS.
Clone Video — 720P and FHD work on both Ultra and UltraS; new sessions start on UltraS 720P.
Pick a higher quality to unlock higher resolutions.
What does Duration do, and how is it tied to quality?
Duration sets how long a generated video is. Like resolution, some duration options are only available at certain quality tiers — unavailable ones are disabled with an "Only available for …" tooltip. When a mode supports long clips, a 10+ button opens a menu of extended durations; options not allowed at your current quality stay disabled there too. Longer durations cost more, since many video modes are priced per second.
For Image to Video and Text to Video, Ultra and UltraS support 5s, 10s, 15s, 20s, 25s, and 30s. Fast remains limited to 4s, 5s, and 10s. In Extend Video, Fast also supports 4s, 5s, and 10s, while Ultra and UltraS support 5s, 10s, 15s, 20s, and 25s.
Why does my duration only allow a few shots?
In multi-shot modes, the number of shots a clip can hold depends on its duration. You'll see hints like "{seconds} second videos support up to {shots} shots. For more shots, increase the video duration." Bump the duration up to fit more shots.
What happens when I change quality and my current settings no longer fit?
The workstation keeps your selections valid. If a setting becomes incompatible with a new quality, it gets adjusted, and you may see a notice like "Some options are no longer available." or "All options are no longer available." For bigger changes — like quality tiers that don't support Elements — you get a confirmation dialog first.
What's the "Some elements will be removed" dialog?
If switching quality would drop any of your filled Elements, the workstation asks first. The dialog reads: "[quality] supports up to [count] elements. Switching now will remove [N] extra elements and their mentions from the prompt." Choose Continue to switch — the overflow Elements (and their @-mentions in your prompt) are removed while everything else you typed stays — or Cancel to keep your current quality.
The same dialog covers both cases: switching to a tier that doesn't support Elements at all (every Element is removed) and lowering to a tier with a smaller Element limit (only the extras beyond the new limit are removed; the first slots survive). It appears in every mode that supports Elements — Generate Image, Image to Video, and Text to Video — on desktop and mobile alike.
Do these controls change the price?
Yes — higher quality, higher resolution, and longer duration all push the cost up, and the number on the Generate button updates live as you change them. See Generate & cost.
On mobile
Quality, resolution, and duration appear as selects in the expanded workstation's action area instead of side-by-side chips. The same availability rules and confirmation dialogs apply. Unavailable resolution options are disabled with the same "Only available for …" hints, while the Image to Video, Text to Video, and Extend Video duration selects filter out durations that the current quality cannot use.