Hardware › Surface Laptop Ultra (RTX Spark N1X, 6144-core GPU) 128GB
Surface Laptop Ultra (RTX Spark N1X, 6144-core GPU) 128GB for local AI: price in Australia & what it can run
A$9,649cheapest found · Microsoft Store AU
128 GBunified (96.0 GB usable)
301 GB/smemory bandwidth
59 tok/sestimated, Llama 2 7B Q4_0
View at Microsoft Store AU → Price checked 19 Oct 2026, 8:18 pm AEST.
New for Oct 2026: NVIDIA's RTX Spark laptop chip (same family as DGX Spark's GB10) with ~301 GB/s unified memory. Windows on Arm; how much of the unified memory Windows lets the GPU use isn't published for every config, so our 75% figure is an assumption. No public llama.cpp benchmark yet: speeds are estimates.
Benchmark
No public benchmark found for this exact hardware. Speeds on this page are bandwidth-based estimates using a median efficiency of 74%.
Specs: https://www.nvidia.com/en-us/products/rtx-spark/specifications/
AU prices found
| Retailer | Listing | Price | Stock | Checked |
|---|---|---|---|---|
Microsoft Store AU | A$9,649 | in stock | 19 Oct 2026, 8:18 pm AEST |
Community measured
No reader results yet. Be the first to submit one →
Which models fit? (8k context)
| Model | Quant | Needs / usable | Fit | Est. tok/s |
|---|---|---|---|---|
Q4_K_M | 464.4 / 96.0 GB | no | – | |
Q4_K_M | 499.4 / 96.0 GB | no | – | |
Q4_K_M | 20.0 / 96.0 GB | fits | 123 estimate | |
Q4_K_M | 348.5 / 96.0 GB | no | – | |
Q4_K_M | 15.7 / 96.0 GB | fits | 184 estimate | |
Q4_K_M | 45.9 / 96.0 GB | fits | 41 estimate | |
Q4_K_M | 22.5 / 96.0 GB | fits | 123 estimate | |
Q4_K_M | 305.3 / 96.0 GB | no | – | |
MiniCPM V 4.7 1B new | Q4_K_M | 2.0 / 96.0 GB | fits | 237 estimate |
Q4_K_M | 464.4 / 96.0 GB | no | – | |
Q4_K_M | 195.9 / 96.0 GB | no | – | |
GLM 5.3 (MoE) new | Q4_K_M | 458.4 / 96.0 GB | no | – |
Q4_K_M | 110.3 / 96.0 GB | no | – | |
Q4_K_M | 18.4 / 96.0 GB | fits | 13 estimate | |
Q4_K_M | 20.2 / 96.0 GB | fits | 123 estimate | |
Q4_K_M | 9.1 / 96.0 GB | fits | 31 estimate | |
Q4_K_M | 23.0 / 96.0 GB | fits | 123 estimate | |
Q4_K_M | 81.4 / 96.0 GB | fits | 3 estimate | |
Q4_K_M | 22.1 / 96.0 GB | fits | 12 estimate | |
Q4_K_M | 17.2 / 96.0 GB | fits | 97 estimate | |
Q4_K_M | 6.0 / 96.0 GB | fits | 46 estimate | |
Q4_K_M | 7.1 / 96.0 GB | fits | 38 estimate | |
Q4_K_M | 4.1 / 96.0 GB | fits | 79 estimate | |
Q4_K_M | 77.0 / 96.0 GB | fits | 37 estimate | |
Q4_K_M | 20.4 / 96.0 GB | fits | 123 estimate | |
gpt-oss 20B (MoE) older | MXFP4 | 12.4 / 96.0 GB | fits | 117 estimate |
gpt-oss 120B (MoE) older | MXFP4 | 63.5 / 96.0 GB | fits | 82 estimate |
Qwen 3 4B older | Q4_K_M | 4.6 / 96.0 GB | fits | 91 estimate |
Qwen 3 8B older | Q4_K_M | 7.2 / 96.0 GB | fits | 45 estimate |
Qwen 3 32B older | Q4_K_M | 23.0 / 96.0 GB | fits | 11 estimate |
Llama 3.1 8B Instruct older | Q4_K_M | 6.9 / 96.0 GB | fits | 46 estimate |
Mistral 7B Instruct v0.3 older | Q4_K_M | 6.5 / 96.0 GB | fits | 51 estimate |