中

Home / GPU ladder / RTX 4080 Super 16GB

RTX 4080 Super 16GB for MiniMax H3: Speed, VRAM & Recommended Setup

16GB GDDR6X · bandwidth 736 GB/s · Ada · sm_89 · estimated with 32GB RAM / fast-disk · updated 2026-09-28

Verdict: Recommended: 1920 × 1088 · 15s, ~58 min 32 s per clip (VRAM peak ~15GB). See the table for higher tiers.
Size (16:9)5s8s10s15s
416 × 256
0.1MP · flicker-prone
48 s
VRAM 15G
54 s
VRAM 15G
59 s
VRAM 15G
1 min 10 s
VRAM 15G
864 × 480
0.4MP
1 min 21 s
VRAM 15G
1 min 51 s
VRAM 15G
2 min 16 s
VRAM 15G
3 min 29 s
VRAM 15G
960 × 544
0.5MP
1 min 33 s
VRAM 15G
2 min 17 s
VRAM 15G
2 min 51 s
VRAM 15G
4 min 39 s
VRAM 15G
1152 × 640
0.7MP
2 min 2 s
VRAM 15G
3 min 17 s
VRAM 15G
4 min 18 s
VRAM 15G
7 min 39 s
VRAM 15G
1344 × 768
1.0MP · official max
2 min 51 s
VRAM 15G
5 min 1 s
VRAM 15G
6 min 54 s
VRAM 15G
13 min 32 s
VRAM 15G
1376 × 768
1.0MP · above native
2 min 55 s
VRAM 15G
5 min 11 s
VRAM 15G
7 min 9 s
VRAM 15G
14 min 8 s
VRAM 15G
1920 × 1088
2.0MP · above native
7 min 9 s
VRAM 15G
15 min 47 s
VRAM 15G
24 min 20 s
VRAM 15G
58 min 32 s
VRAM 15G

Times are per whole clip (model load and text encoding included); the first run is slower due to kernel compilation. ⚠ = weight offloading / noticeably slower; ✕ = not recommended.

Reading the numbers

With the recommended combo, the green (no disk spill) ceiling is 1920 × 1088 · 15s at ~58 min 32 s per clip. To unlock 2.0MP and full durations: step up to a 32GB card (e.g. 5090), or use a smaller quant (GGUF Q3/Q4).

Recommended combo (Blackwell-specific (50-series 16GB+))

768p 5s; 4-8 steps; 1-5 min per clip

NVFP4 requires sm_120; SLA / VSA sparse attention can speed things up further (1.4-1.75x on long clips, see the attention options in step 2)

FAQ

Can the RTX 4080 Super 16GB run MiniMax H3?

Yes. Estimated with 32GB RAM + fast-disk: 1920×1088·15s at ~58 min 32 s per clip. Configs that exceed VRAM fall back to weight offloading (slower, but they still render).

How fast is H3 on the RTX 4080 Super 16GB?

About 58 min 32 s per 1920×1088·15s clip (including model load and encoding). Longer or denser clips scale quadratically in the attention term.

How much RAM do I need?

At least 64GB of system RAM recommended (minimum 32GB plus a 16GB pagefile). With 32GB, always enable --fast-disk or weights spill into the pagefile.

How can I make it faster?

Stack step distillation (e.g. LightX2V Turbo 4-step v1.1 768p), enable SageAttention / Kitchen attention; for long clips try SLA / VSA sparse attention. All of it is switchable in step 2 of the calculator.

→ Run the numbers for your RAM and budget in the calculator (RTX 4080 Super 16GB pre-filled)

Same tier: RTX 3090 24GB · RTX 4070 Ti Super 16GB · RTX 4080 16GB · RTX 4080 32GB (modded) · RTX 4080 Super 32GB (modded) · RTX 4090 24GB · RTX 4090 D 24GB · RTX 4090 48GB (modded)

Related: Hardware guide · Acceleration compared · 8GB VRAM setups