Models › Llama 3.1 8B Instruct
Can I run Llama 3.1 8B Instruct in Australia?
Yes. The cheapest hardware we track that holds Llama 3.1 8B Instruct at Q4_K_M with an 8k context is the Intel Arc B580 12GB at A$419 (Scorptec, checked 20 Oct 2026, 7:44 am AEST). Expect roughly 55 tokens/sec (estimate).
Memory needed by quantisation (8k context)
| Quant | Weights | Total needed | Cheapest hardware that fits |
|---|---|---|---|
Q4_K_M | 4.9 GB | 6.9 GB | Intel Arc B580 12GB · A$419 |
Q5_K_M | 5.7 GB | 7.8 GB | Intel Arc B580 12GB · A$419 |
Q6_K | 6.6 GB | 8.7 GB | Intel Arc B580 12GB · A$419 |
Q8_0 | 8.5 GB | 10.6 GB | Intel Arc B580 12GB · A$419 |
F16 | 16.1 GB | 18.1 GB | AMD Radeon AI PRO R9700 32GB · A$2,299 |
All hardware that fits at Q4_K_M
| Hardware | Needs / usable | Est. tok/s | Price | Retailer link |
|---|---|---|---|---|
6.9 / 11.5 GB fits | 55 estimate | A$419 Scorptec checked 20 Oct 2026, 7:44 am AEST | ||
6.9 / 11.5 GB fits | 59 estimate | A$539 Scorptec checked 20 Oct 2026, 7:38 am AEST | ||
6.9 / 15.5 GB fits | 53 estimate | A$699 Mwave checked 20 Oct 2026, 7:38 am AEST | ||
6.9 / 15.5 GB fits | 71 estimate | A$1,099 Scorptec checked 20 Oct 2026, 7:39 am AEST | ||
6.9 / 15.5 GB fits | 93 estimate | A$1,099 Mwave checked 20 Oct 2026, 7:38 am AEST | ||
6.9 / 15.5 GB fits | 107 estimate | A$1,169 Mwave checked 20 Oct 2026, 7:38 am AEST | ||
6.9 / 11.5 GB fits | 99 estimate | A$1,249 Scorptec checked 20 Oct 2026, 7:39 am AEST | ||
6.9 / 12.0 GB fits | 23 estimate | A$1,449 Apple AU checked 20 Oct 2026, 7:44 am AEST | ||
6.9 / 12.0 GB fits | 19 estimate | A$1,597 Mwave checked 20 Oct 2026, 7:44 am AEST | ||
6.9 / 15.5 GB fits | 138 estimate | A$1,699 Scorptec checked 20 Oct 2026, 7:40 am AEST | ||
6.9 / 15.5 GB fits | 142 estimate | A$2,299 Scorptec checked 20 Oct 2026, 7:40 am AEST | ||
6.9 / 31.5 GB fits | 113 estimate | A$2,299 Mwave checked 20 Oct 2026, 7:38 am AEST | ||
6.9 / 18.0 GB fits | 53 estimate | A$2,699 Apple AU checked 20 Oct 2026, 7:44 am AEST | ||
6.9 / 18.0 GB fits | 52 estimate | A$2,999 Apple AU checked 20 Oct 2026, 7:44 am AEST | ||
6.9 / 23.5 GB fits | 133 estimate | A$3,599 Scorptec · not in stock online checked 20 Oct 2026, 7:41 am AEST | ||
6.9 / 48.0 GB fits | 39 estimate | A$3,915 Minisforum AU · not in stock online checked 20 Oct 2026, 7:44 am AEST | ||
6.9 / 27.0 GB fits | 70 estimate | A$4,299 Apple AU checked 20 Oct 2026, 7:44 am AEST | ||
6.9 / 96.0 GB fits | 39 estimate | A$5,719 Minisforum AU checked 20 Oct 2026, 7:44 am AEST | ||
6.9 / 48.0 GB fits | 39 estimate | A$5,894 Mwave checked 20 Oct 2026, 7:44 am AEST | ||
6.9 / 48.0 GB fits | 46 estimate | A$7,149 Microsoft Store AU · stale checked 19 Oct 2026, 8:18 pm AEST | ||
6.9 / 96.0 GB fits | 39 estimate | A$8,999 Mwave · not in stock online checked 20 Oct 2026, 7:44 am AEST | ||
6.9 / 96.0 GB fits | 45 estimate | A$8,999 Mwave checked 20 Oct 2026, 7:44 am AEST | ||
6.9 / 72.0 GB fits | 183 estimate | A$9,499 Apple AU checked 20 Oct 2026, 7:44 am AEST | ||
6.9 / 96.0 GB fits | 46 estimate | A$9,649 Microsoft Store AU · stale checked 19 Oct 2026, 8:18 pm AEST | ||
6.9 / 31.5 GB fits | 226 estimate | A$10,999 Scorptec checked 20 Oct 2026, 7:41 am AEST | ||
6.9 / 192.0 GB fits | 139 estimate | A$11,449 Apple AU checked 20 Oct 2026, 7:44 am AEST | ||
6.9 / 95.5 GB fits | 214 estimate | A$27,599 Scorptec checked 20 Oct 2026, 7:42 am AEST | ||
6.9 / 15.5 GB fits | 44 estimate | price unavailable | ||
6.9 / 23.5 GB fits | 123 estimate | price unavailable | ||
6.9 / 23.5 GB fits | 145 estimate | price unavailable | ||
6.9 / 23.5 GB fits | 142 estimate | price unavailable | ||
6.9 / 36.0 GB fits | 93 estimate | price unavailable | ||
6.9 / 48.0 GB fits | 39 estimate | price unavailable | ||
6.9 / 96.0 GB fits | 39 estimate | price unavailable |
The tok/s figures are bandwidth-based estimates for this model. Where a device has a measured benchmark, the estimate is calibrated to that run. They are not measurements. Methodology.
Graphics card + system RAM (partial layer offload)
Dense models can split layers between GPU and system RAM, but every token reads all the weights, so the RAM part is slow. All figures below are estimates (Q4_K_M, 8k context, dual-channel DDR5-5600). How we estimate.
| Graphics card | Card price | GPU only | + 32GB RAM | + 64GB RAM | + 128GB RAM |
|---|---|---|---|---|---|
A$419 | fits ~55 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
A$539 | fits ~59 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
A$699 | fits ~53 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
A$1,099 | fits ~71 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
A$1,099 | fits ~93 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
A$1,169 | fits ~107 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
A$1,249 | fits ~99 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
A$1,699 | fits ~138 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
A$2,299 | fits ~142 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
A$2,299 | fits ~113 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
A$3,599 | fits ~133 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
A$10,999 | fits ~226 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
A$27,599 | fits ~214 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
n/a | fits ~44 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
n/a | fits ~123 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
n/a | fits ~145 tok/s | fits in VRAM | fits in VRAM | fits in VRAM | |
n/a | fits ~142 tok/s | fits in VRAM | fits in VRAM | fits in VRAM |
DDR5 kit prices today: 32GB: A$610 (Mwave, checked 20 Oct 2026, 7:44 am AEST) · 64GB: A$1,499 (Mwave, checked 20 Oct 2026, 7:44 am AEST) · 128GB: A$2,511 (Mwave, not in stock online, checked 20 Oct 2026, 7:44 am AEST). Card + kit only; the rest of the PC isn't included. RAM-offload speeds are estimates and can be optimistic, so check community results below.
Community measured
No reader results yet. Be the first to submit one →
Try other quantisations and context lengths in the Can I run it? tool → · Submit your own Llama 3.1 8B Instruct benchmark →