Models / Llama 3.1 8B Instruct / Calculator

Llama 3.1 8B Instruct cost & VRAM calculator

What does Llama 3.1 8B Instruct cost to run for your workload, and can you run it on your own hardware? Set your workload below - we compute per-provider API cost live and tell you honestly whether local hardware can run it.

Your workload

Casual Power user Team of 5 Codebase task

API cost for your workload

Provider Rate ($/1M) Monthly cost
DeepInfra $0.02 in · $0.04 out · $0.02 cache $0.0 cheapest
Novita $0.02 in · $0.05 out · $0.02 cache $0.0 cheapest
Groq $0.05 in · $0.08 out · $0.02 cache $0.0 cheapest
Cloudflare $0.15 in · $0.29 out · $0.02 cache $0.0 cheapest
Together AI may be stale $0.18 in · $0.18 out $0.0 cheapest
Fireworks AI may be stale $0.20 in · $0.20 out $0.0 cheapest
CoreWeave $0.22 in · $0.22 out · $0.22 cache $0.01

Monthly cost is an estimate from list prices and your workload - verify against the provider before committing. Cached fraction applies the cache rate to that share of input.

Can you run it locally?

YES - IT RUNS ON HARDWARE YOU CAN BUY

Llama 3.1 8B Instruct fits on individual rigs. Here's the hardware that runs it comfortably; the full per-quant VRAM matrix and tok/s estimates live on the model page.

4x H100 80GB (320GB) 320GB Q4_K_M
1637.1t/s
NVIDIA DGX Station 748GB 748GB Q4_K_M
977.3t/s
8x RTX 3090 rack (192GB) 192GB Q4_K_M
915.0t/s
4x RTX 5090 (128GB) 128GB Q4_K_M
875.7t/s
AMD Instinct MI300X (192GB) 192GB Q4_K_M
650.5t/s
4x RTX 4090 (96GB) 96GB Q4_K_M
492.6t/s
2x RTX 5090 (64GB) 64GB Q4_K_M
437.9t/s
2x RTX 3090 (48GB) 48GB Q4_K_M
228.8t/s
Single RTX 5090 (32GB) 32GB Q4_K_M
218.9t/s
RTX PRO 6000 Blackwell (96GB) 96GB Q4_K_M
218.9t/s
Mac Studio M4 Ultra 192GB 192GB Q4_K_M
145.5t/s
Mac Studio M4 Ultra 512GB 512GB Q4_K_M
145.5t/s
Single RTX 4090 (24GB) 24GB Q4_K_M
123.2t/s
MacBook Pro M5 Max 128GB 128GB Q4_K_M
81.8t/s
Single GTX 1080 Ti (11GB) 11GB Q4_K_M
59.1t/s
Dual EPYC 9004 + 768GB DDR5-4800 768GB Q4_K_M
56.3t/s
DGX Spark 128GB unified 128GB Q4_K_M
33.4t/s
Ryzen AI Max+ 395 128GB 128GB Q4_K_M
31.3t/s
Jetson AGX Orin 64GB 64GB Q4_K_M
25.0t/s
Epyc + 512GB DDR4-3200 + 2x RTX 3090 560GB Q4_K_M
25.0t/s
Epyc + 512GB DDR4-2400 + 2x RTX 3090 560GB Q4_K_M
18.8t/s
NVIDIA Jetson Orin NX 16GB 16GB Q4_K_M
12.5t/s
Q4_K_M
4.5GB weights 4.5GB min 6.0GB rec
Q8_0
8.0GB weights 8.0GB min 9.0GB rec
FP16
16.0GB weights 16.0GB min 18.0GB rec

Want the own-vs-rent payback (capex vs your monthly API bill)? The budget tool computes that against representative rigs.

Full model card API pricing table Generic token calculator