google/gemma-4-26B-A4B-it
google/gemma-4-26B-A4B-it
Parameters
4B
Context
256K
Downloads
10.1M
Likes
1.4K
Architecture
gemma4
30 layers · 8 KV heads · head 256 · hidden 2816
GPU Requirements
VRAM / RAM / disk estimates by quantization and usage scenario
| Quantization | Minimum | Recommended | Production |
|---|---|---|---|
| FP1616bit | 14.1 GB RAM 32.0 · Disk 50.7 | 35.4 GB RAM 53.1 · Disk 76.7 | 181.5 GB RAM 272.2 · Disk 115.7 |
| FP88bit | 9.4 GB RAM 32.0 · Disk 44.8 | 30.4 GB RAM 45.7 · Disk 70.8 | 176.3 GB RAM 264.5 · Disk 109.8 |
| INT88bit | 9.4 GB RAM 32.0 · Disk 44.8 | 30.4 GB RAM 45.7 · Disk 70.8 | 176.3 GB RAM 264.5 · Disk 109.8 |
| AWQ4bit | 7.1 GB RAM 32.0 · Disk 42.0 | 28.1 GB RAM 42.1 · Disk 68.0 | 173.8 GB RAM 260.8 · Disk 107.0 |
| GPTQ4bit | 7.1 GB RAM 32.0 · Disk 42.1 | 28.1 GB RAM 42.1 · Disk 68.1 | 173.9 GB RAM 260.8 · Disk 107.1 |
| GGUF4bit | 7.2 GB RAM 32.0 · Disk 42.1 | 28.1 GB RAM 42.2 · Disk 68.1 | 173.9 GB RAM 260.9 · Disk 107.1 |
GPU Recommendations
1× H100 SXM
80 GB VRAM
≈ 1726.8 tok/s
Effective production throughput ≈ 1208.8 tok/s
score 0.583
2× RTX 4070S Ti
32 GB VRAM
≈ 0 tok/s (N/A — benchmark unavailable)
Effective production throughput ≈ 0 tok/s
score 0.6117
8× H200 SXM
1128 GB VRAM
≈ 7917.5 tok/s
Effective production throughput ≈ 5542.3 tok/s
score 0.2197
1× RTX A6000
48 GB VRAM
≈ 395.9 tok/s
Effective production throughput ≈ 277.1 tok/s
score 0.6469
💬 Community — real deployment experience
0 Articles · 0 Benchmarks
GPUs that fit (single card)
RTX A6000
48 GB VRAM
$210
/mo · vast
RTX 5090
32 GB VRAM
$234
/mo · vast
L40S
48 GB VRAM
$341
/mo · vast
RTX PRO 6000 WS
48 GB VRAM
$482
/mo · vast
RTX PRO 6000 S
48 GB VRAM
$633
/mo · vast