OCR / VisionMIT
Pixtral Large 124B
mistralai/Pixtral-Large-Instruct-2411
Parameters
124B
Context
128K
Downloads
59
Likes
434
Architecture
transformer
GPU Requirements
VRAM / RAM / disk estimates by quantization and usage scenario
| Quantization | Minimum | Recommended | Production |
|---|---|---|---|
| FP1616bit | 295.8 GB RAM 443.6 · Disk 399.3 | 311.2 GB RAM 466.8 · Disk 425.3 | 340.3 GB RAM 510.5 · Disk 464.3 |
| FP88bit | 149.7 GB RAM 224.5 · Disk 219.2 | 158.5 GB RAM 237.7 · Disk 245.2 | 181.0 GB RAM 271.5 · Disk 284.2 |
| INT88bit | 149.7 GB RAM 224.5 · Disk 219.2 | 158.5 GB RAM 237.7 · Disk 245.2 | 181.0 GB RAM 271.5 · Disk 284.2 |
| AWQ4bit | 79.4 GB RAM 119.0 · Disk 132.5 | 85.0 GB RAM 127.5 · Disk 158.5 | 104.3 GB RAM 156.4 · Disk 197.5 |
| GPTQ4bit | 80.3 GB RAM 120.4 · Disk 133.6 | 85.9 GB RAM 128.9 · Disk 159.6 | 105.3 GB RAM 157.9 · Disk 198.6 |
| GGUF4bit | 82.1 GB RAM 123.2 · Disk 135.8 | 87.8 GB RAM 131.8 · Disk 161.8 | 107.3 GB RAM 160.9 · Disk 200.8 |
GPU Recommendations
Best ValueBEST
1× H200 SXM
141 GB VRAM
$2,555/mo
≈ 80.1 tok/s
Effective production throughput ≈ 56.1 tok/s
score 0.6318
Cheapest
8× Tesla V100
128 GB VRAM
$162/mo
≈ 0 tok/s (N/A — benchmark unavailable)
Effective production throughput ≈ 0 tok/s
score 0.5337
Performance
8× H200 SXM
1128 GB VRAM
$20,440/mo
≈ 256.4 tok/s
Effective production throughput ≈ 179.5 tok/s
score 0.2197
Min Complexity
2× RTX A6000
96 GB VRAM
$420/mo
≈ 20.5 tok/s
Effective production throughput ≈ 14.4 tok/s
score 0.6115
💬 Community — real deployment experience
0 Articles · 0 Benchmarks
GPUs that fit (single card)
H200 SXM
141 GB VRAM
$2,555
/mo · lambda
B200
180 GB VRAM
$3,013
/mo · vast
B200 SXM
180 GB VRAM
—