MiniMax H3 Image
Parameters
33.1B
Modality
Text
Architecture
Dense
License
other
Release Date
3 Nov 2026
API Pricing (per 1M)
Self-hosted only
Hardware & System Requirements
Precision
Workload Memory: 23.38 GB
INT4 · 1024x1024
DiT Backbone Weights: 15.41 GB
VAE Decode Workspace: 5.47 GB
Runtime & Graphs: 2.50 GB
Minimum Required Hardware
1x NVIDIA A40 48GB
23.4 / 43.2 GB usable
2x RTX 4090 24GB
23.4 / 43.2 GB usable · TP 2
Apple M3 Pro 36GB
23.4 / 27.0 GB usable
Determine requirements for your specific workload:
Open in VRAM Calculator
Runtime requirements
Checkpoint format
FP16
CUDA
>= 11.8
PyTorch
>= 2.1
Transformers
>= 4.40.1
| Architecture | Precision | Attention kernel |
|---|---|---|
Blackwell, Hopper | Native FP16 | FlashAttention-3 |
Ada Lovelace, Ampere | Native FP16 | FlashAttention-2 |
Architecture Diagram
About MiniMax H3 Image
MiniMax H3 Image is a visual generation foundation model developed by MiniMax AI supporting high-fidelity text-to-image synthesis. It provides detailed prompt comprehension and compositional rendering.
Specifications
Attention
Attention Structure
Multi-Head Attention
Attention Heads
56
Key-Value Heads
56
Attention Head Dimension
128
Position Embedding
ROPE
Normalization
RMS Normalization
Activation Function
GELU
Dimensions
Base Precision
FP16
Hidden Dimension Size
5,376
Number of Layers
50
FFN Intermediate Size (Dense)
14,336
Resources
About MiniMax M3
MiniMax's flagship M3 model family, released June 1, 2026, is powered by MiniMax Sparse Attention (MSA) architecture, offering 1M context capabilities at exceptionally low compute cost and optimized for long-horizon agentic workflows.