Gemma 2 9B hardware requirements
9.24B parameters. Needs about 9 GB at Q4_K_M with 8K context. Model card.
Memory needed at each quantization
| Quant | Quality | 8K context | 32K context |
|---|---|---|---|
| FP16 | Full quality | 21.9 GB | 30.4 GB |
| Q8_0 | Near-lossless | 13.2 GB | 21.7 GB |
| Q6_K | Excellent | 11 GB | 19.5 GB |
| Q5_K_M | Very good | 10 GB | 18.4 GB |
| Q4_K_M | Good, most popular | 9 GB | 17.5 GB |
| Q3_K_M | Noticeably degraded | 7.9 GB | 16.4 GB |
| Q2_K | Heavily degraded | 6.9 GB | 15.3 GB |
Run it with Ollama: ollama run gemma2:9b (downloads Q4_K_M by default).
Run it in the cloud
Don't want to buy hardware? Run Gemma 2 9B on a rented GPU today.
Rent a cloud GPU by the hour. No upfront cost, cancel any time.
Every GPU and Mac checked
| GPU | Runs? | Best fit |
|---|---|---|
| NVIDIA RTX 3060 12GB | Yes | Q6_K · ~11 GB · ~29 tok/s |
| NVIDIA RTX 3070 | Yes, with heavy quantization | Q3_K_M · ~7.9 GB · ~60 tok/s |
| NVIDIA RTX 3080 10GB | Yes | Q5_K_M · ~10 GB · ~69 tok/s |
| NVIDIA RTX 3090 | Yes | FP16 · ~21.9 GB · ~30 tok/s |
| NVIDIA RTX 4060 | Yes, with heavy quantization | Q3_K_M · ~7.9 GB · ~36 tok/s |
| NVIDIA RTX 4060 Ti 16GB | Yes | Q8_0 · ~13.2 GB · ~18 tok/s |
| NVIDIA RTX 5060 Ti 16GB | Yes | Q8_0 · ~13.2 GB · ~27 tok/s |
| NVIDIA RTX 4070 | Yes | Q6_K · ~11 GB · ~40 tok/s |
| NVIDIA RTX 4070 Ti Super | Yes | Q8_0 · ~13.2 GB · ~41 tok/s |
| NVIDIA RTX 4080 | Yes | Q8_0 · ~13.2 GB · ~44 tok/s |
| NVIDIA RTX 4090 | Yes | FP16 · ~21.9 GB · ~33 tok/s |
| NVIDIA RTX 5070 | Yes | Q6_K · ~11 GB · ~53 tok/s |
| NVIDIA RTX 5070 Ti | Yes | Q8_0 · ~13.2 GB · ~55 tok/s |
| NVIDIA RTX 5080 | Yes | Q8_0 · ~13.2 GB · ~59 tok/s |
| NVIDIA RTX 5090 | Yes | FP16 · ~21.9 GB · ~58 tok/s |
| AMD Radeon RX 7800 XT | Yes | Q8_0 · ~13.2 GB · ~38 tok/s |
| AMD Radeon RX 7900 XTX | Yes | FP16 · ~21.9 GB · ~31 tok/s |
| Apple M4 (16GB) | Yes | Q5_K_M · ~10 GB · ~11 tok/s |
| Apple M1 (16GB) | Yes | Q5_K_M · ~10 GB · ~6 tok/s |
| Apple M2 Pro (32GB) | Yes | Q8_0 · ~13.2 GB · ~12 tok/s |
| Apple M3 Max (64GB) | Yes | FP16 · ~21.9 GB · ~13 tok/s |
| Apple M4 Pro (48GB) | Yes | FP16 · ~21.9 GB · ~9 tok/s |
| Apple M4 Max (128GB) | Yes | FP16 · ~21.9 GB · ~18 tok/s |
| Apple M2 Ultra (192GB) | Yes | FP16 · ~21.9 GB · ~26 tok/s |