Mistral Large 4 1T A52B
Active Parameters
1T
Context Length
1.05M
Modality
Text
Architecture
Mixture of Experts (MoE)
License
other
Release Date
30 Sept 2026
Knowledge Cutoff
-
API Pricing (per 1M)
Self-hosted only
System Requirements
VRAM requirements for different quantization methods and context sizes
1,024 tokens
Consumer
128x RTX 4090
24GB VRAM
Datacenter
32x NVIDIA A100
80GB VRAM
Apple Silicon
28x Apple M3 Max
128GB VRAM
1,048,576 tokens
Consumer
233x RTX 4090
24GB VRAM
Datacenter
54x NVIDIA A100
80GB VRAM
Apple Silicon
51x Apple M3 Max
128GB VRAM
Architecture Diagram
Evaluation Benchmarks
No evaluation benchmarks for Mistral Large 4 1T A52B available.
Rankings
Overall Rank
-
Coding Rank
-
About Mistral Large 4 1T A52B
Mistral Large 4 1T A52B is a trillion-parameter mixture-of-experts model activating 52 billion parameters per token. It delivers frontier-level reasoning, multilingual fluency, and coding capabilities.
Technical Specifications
Attention
Attention Structure
Compressed Sparse Attention 2
Attention Heads
64
Key-Value Heads
64
Attention Head Dimension
64
Position Embedding
ROPE
RoPE Theta
10,000
KV Source Layers
40
KV Latent Dimension
576
Indexer Key Dimension
128
Normalization
RMS Normalization
Activation Function
SwigLU
Dimensions
Hidden Dimension Size
6,144
Number of Layers
72
FFN Intermediate Size (Dense)
24,576
Tokenizer
Vocabulary Size
204,800
Mixture of Experts
Total Expert Parameters
52.0B
Number of Experts
128
Active Experts
4
Shared Experts
1
FFN Intermediate Size (per Expert)
6,144
Dense Layers Before MoE
1
Resources
About Mistral Large 4
The Mistral Large 4 model family developed by Mistral AI.