MiniMax · Video · 33.1B parameters · MiniMax H3 Community License Agreement
MiniMax H3
Needs 40 GB at FP8 / INT8. The smallest real configuration today is 2× Tesla V100 32GB at Vast.ai, $0.14 per hour, about $99 a month at 730 hours.
Parameters33.1Bdense
CategoryVideosized on weights + headroom
Checkpoint498 GBBF16, as published
LicenseMiniMax H3 Community License Agreementcheck before commercial use
ReleasedJul 28, 2026documented date
Smallest configuration · FP8 / INT8
2× Tesla V100 32GB64 GB VRAM · 62% used
- Per hour
- $0.14
- Per month
- $99
- Cheapest at
Vast.ai
Memory · 3 precisionsHow much VRAM it needs
Weights plus 20% runtime overhead. For video models, activations also depend on resolution, clip length and batch size; the per-GPU estimates below give the practical configuration.
| Precision | Weights | Weights + headroom |
|---|---|---|
| BF16published checkpoint | 498 GB | 598 GB |
| FP8 / INT8estimated | 33 GB | 40 GB |
| INT4 / MXFP4estimated | 20 GB | 24 GB |
Cross-checked on Hugging Face Oct 28, 2026: 33.1B parameters in the safetensors headers, 498.3 GB of weight files. Method: weights as published at the native precision, verified on Hugging Face 2026-10-28; other precisions estimated at 1 (8-bit) and 0.6 (4-bit) bytes per parameter; × 1.2 runtime headroom; KV cache not added (architecture not readable).
Smallest configuration per GPU · 109 GPUs · cheapest eligible on-demand offer per GPU count, every provider on fileWhat it costs to run, GPU by GPU
The fewest GPUs of one type whose total memory fits the model at short context, priced from the cheapest eligible on-demand offer for exactly that count (45-day window, newest reading per provider), with the next providers beside it. Monthly = 730 hours.
| GPU | Smallest config | VRAM used | $ / hour | $ / month | Cheapest at |
|---|---|---|---|---|---|
| Tesla V100 32GBBest fit | 2× 32 GB64 GB total | 62% | $0.14 | $99 | Vast.aiTrust 78 · read Nov 2 · receipt · also Cyfuture AI $1.13, io.net $2.52 |
| RTX 3060 | 4× 12 GB48 GB total | 83% | $0.16 | $118 |  Clore.aiTrust 49 · read Nov 2 · receipt · also Vast.ai $0.21, Edgevana $0.31 |
| Tesla P100 | 4× 16 GB64 GB total | 62% | $0.19 | $142 | Vast.aiTrust 78 · read Nov 2 · receipt · also Edgevana $0.43 |
| RTX 3090 | 2× 24 GB48 GB total | 83% | $0.21 | $157 | Vast.aiTrust 78 · read Nov 2 · receipt · also Clore.ai $0.24, Lium $0.44 |
| Quadro RTX 8000 | 1× 48 GB48 GB total | 83% | $0.23 | $167 | Vast.aiTrust 78 · read Nov 2 · receipt · also TensorDock $0.32 |
| Titan RTX | 2× 24 GB48 GB total | 83% | $0.27 | $196 | Vast.aiTrust 78 · read Nov 2 · receipt · also Edgevana $0.52 |
| Titan Xp | 4× 12 GB48 GB total | 83% | $0.27 | $196 | Vast.aiTrust 78 · read Nov 2 · receipt |
| L40 | 1× 48 GB48 GB total | 83% | $0.27 | $196 | Vast.aiTrust 78 · read Nov 2 · receipt · also io.net $0.66, Runpod $0.69 |
| RTX 2080 Ti | 4× 11 GB44 GB total | 90% tight | $0.28 | $202 | Vast.aiTrust 78 · read Nov 2 · receipt |
| RTX A6000 | 1× 48 GB48 GB total | 83% | $0.33 | $241 |  GPURentonot scored · read Nov 2 · receipt · also gcube $0.34, Thunder Compute $0.35 |
| RTX 4000 Ada | 2× 20 GB40 GB total | 99% tight | $0.35 | $254 | Vast.aiTrust 78 · read Nov 2 · receipt · also iwinv $0.57, Akamai Cloud $1.05 |
| A40 | 1× 48 GB48 GB total | 83% | $0.35 | $255 | GPURentonot scored · read Nov 2 · receipt · also Vast.ai $0.48, Exoscale $1.05 |
| RTX 6000 | 2× 24 GB48 GB total | 83% | $0.35 | $256 | Vast.aiTrust 78 · read Nov 2 · receipt · also Edgevana $0.40 |
| RTX 5060 Ti | 4× 16 GB64 GB total | 62% | $0.35 | $258 | Clore.aiTrust 49 · read Nov 2 · receipt · also Vast.ai $0.40, Edgevana $0.68 |
| RTX A4000 | 4× 16 GB64 GB total | 62% | $0.36 | $266 | Vast.aiTrust 78 · read Nov 2 · receipt · also AtmosCompute $0.66, Spheron $3.84 |
| RTX 3080 | 4× 10 GB40 GB total | 99% tight | $0.38 | $274 | Clore.aiTrust 49 · read Nov 2 · receipt · also Vast.ai $0.51, Edgevana $1.23 |
| RTX A5000 | 2× 24 GB48 GB total | 83% | $0.38 | $274 | Vast.aiTrust 78 · read Nov 2 · receipt · also Edgevana $0.74, Airon AI $0.90 |
| RTX 4070 Super | 4× 12 GB48 GB total | 83% | $0.38 | $274 | Vast.aiTrust 78 · read Nov 2 · receipt · also Clore.ai $0.58, AtmosCompute $0.69 |
| Tesla V100 16GB | 4× 16 GB64 GB total | 62% | $0.38 | $275 | Vast.aiTrust 78 · read Nov 2 · receipt · also CloudRift $1.00, AtmosCompute $1.30 |
| A100 40GB | 1× 40 GB40 GB total | 99% tight | $0.39 | $283 | Vast.aiTrust 78 · read Nov 2 · receipt · also Edgevana $0.58, AtmosCompute $0.73 |
| RTX 6000 Ada | 1× 48 GB48 GB total | 83% | $0.40 | $290 | |
| RTX PRO 5000 | 1× 48 GB48 GB total | 83% | $0.40 | $294 | Vast.aiTrust 78 · read Nov 2 · receipt · also Edgevana $0.57, Clore.ai $0.63 |
| Quadro RTX 5000 | 4× 16 GB64 GB total | 62% | $0.40 | $294 | Vast.aiTrust 78 · read Nov 2 · receipt |
| RTX 3090 Ti | 2× 24 GB48 GB total | 83% | $0.42 | $304 | Clore.aiTrust 49 · read Nov 2 · receipt · also Vast.ai $0.48, Edgevana $0.59 |
| The 24 cheapest of 72 GPU types that fit. Every GPU, any precision and context, in the finder | |||||
Throughput estimates · Oct 4, 2026Estimated speed, in min per clip
Video. Speed estimated from memory bandwidth and model size, never measured. The cost per video second is an estimate at the prices read on Oct 4, 2026; it is not recomputed here.
| GPU | Cards | Precision | min per clip | per video second · estimate |
|---|---|---|---|---|
| RTX 4090 | 4 | FP8 | 110 | $0.278 |
| RTX 5090 | 4 | FP8 | 64.7 | $0.342 |
| RTX PRO 6000 | 1 | FP8 | 120 | $0.381 |
| L40S | 2 | FP8 | 146 | $0.444 |
| H100 SXM | 1 | FP8 | 96.0 | $0.490 |
| H100 PCIe | 1 | FP8 | 132 | $0.492 |
| B200 | 1 | FP8 | 48.0 | $0.550 |
| H200 | 1 | FP8 | 90.0 | $0.619 |
| MI300X | 1 | FP8 | 138 | $0.628 |
| A100 80GB | 2 | BF16 | 122 | $0.638 |
cards, precision and speed estimated from memory bandwidth and model size; cost recomputed by the site from verified prices · cheapest per output at the prices of Oct 4, 2026: RTX 4090Why estimates, and their limits →
MiniMax H3 · related