will it local
NVIDIA's RTX Spark N1X is the GB10-class silicon shipping in Windows machines: up to 20-core Grace CPU, up to 6,144-core Blackwell RTX GPU, up to 128GB LPDDR5X unified. First machines: Microsoft Surface Laptop Ultra (availability from October 16) and Surface RTX Spark Dev Box (November, US-only); partner laptops from ASUS, Dell, HP, Lenovo and MSI follow. No independent tok/s reports yet - treat it as a 128GB capacity-first Windows machine like the DGX Spark, not a 5090-class bandwidth machine.
Shared-memory device — speed scales with RAM bandwidth (273.0 GB/s shared RAM), usable for inference ≈ 22.0 GB of 24 GB system RAM
→explore the full catalogupgrade paths with live prices · what fits on each machinecomputed band 13.0-35.0 tok/s for 8B Q4_K_M (roofline, 273.0 GB/s shared RAM)
Effective decode window: 0.25–0.65 of 273 GB/s nominal → ~68–177 GB/s effective (llama.cpp decode, Q4_K_M basis; card-specific window from cited community benches; nominal is LPDDR/JEDEC channel arithmetic (shared memory))
→ 13b-q4 full 14.6 GB needed of 22 GB usable — headroom for context
→ 3b-q4 full 2.8 GB needed of 22 GB usable — headroom for context
→ 4b-q4 full 3.7 GB needed of 22 GB usable — headroom for context
→ 8b-q4 full 6.1 GB needed of 22 GB usable — headroom for context
no single-card bench published — band is computed (see notes)
Nominal bandwidth: 273.0 GB/s shared memory — see the effective decode window above
→ qwen3:0.6b on Tesla M60 (16GB, used) ($35) 0.6B · 0.6GB · needs ~1GB (4k ctx)
→ qwen3:1.7b on Tesla M60 (16GB, used) ($35) 1.7B · 1.4GB · needs ~1.8GB (4k ctx)
→ qwen3:4b on Tesla M60 (16GB, used) ($35) 4B · 2.6GB · needs ~3.2GB (4k ctx)
→ llama3.2:1b on Tesla M60 (16GB, used) ($35) 1B · 1.3GB · needs ~1.4GB (4k ctx)
→ llama3.2:3b on Tesla M60 (16GB, used) ($35) 3B · 2.0GB · needs ~2.4GB (4k ctx)
→ gemma3:1b on Tesla M60 (16GB, used) ($35) 1B · 1.0GB · needs ~1.1GB (4k ctx)
→ phi4:mini on Tesla M60 (16GB, used) ($35) mini 3.8B · 2.5GB · needs ~3GB (4k ctx)
→ moondream:2b on Tesla M60 (16GB, used) ($35) 2B · 1.7GB · needs ~3.7GB (4k ctx)
→ whisper:tiny on Tesla M60 (16GB, used) ($35) tiny · 0.1GB · needs ~2.1GB (4k ctx)
→ whisper on Tesla M60 (16GB, used) ($35) base · 0.3GB · needs ~2.3GB (4k ctx)
→ whisper:small on Tesla M60 (16GB, used) ($35) small · 0.9GB · needs ~2.9GB (4k ctx)
→ whisper:medium on Tesla M60 (16GB, used) ($35) medium · 3.1GB · needs ~5.1GB (4k ctx)
→ kokoro:82m on Tesla M60 (16GB, used) ($35) 82M · 0.3GB · needs ~2.3GB (4k ctx)
→ gemma3:4b on Tesla M60 (16GB, used) ($35) 4B · 3.3GB · needs ~3.8GB (4k ctx)
→ deepseek-r1:7b on Tesla M60 (16GB, used) ($35) distill 7B · 4.7GB · needs ~4.9GB (4k ctx)
→ mistral:7b on Tesla M60 (16GB, used) ($35) 7B · 4.1GB · needs ~4.6GB (4k ctx)
→ llava:7b on Tesla M60 (16GB, used) ($35) 7B · 4.7GB · needs ~6.7GB (4k ctx)
→ qwen3:8b on Tesla M60 (16GB, used) ($35) 8B · 5.2GB · needs ~5.8GB (4k ctx)
→ qwen2.5vl:7b on Tesla M60 (16GB, used) ($35) 7B · 5.6GB · needs ~5.8GB (4k ctx)
→ llama3.1:8b on Tesla M60 (16GB, used) ($35) 8B · 4.9GB · needs ~5.4GB (4k ctx)
→ deepseek-r1:8b on Tesla M60 (16GB, used) ($35) distill 8B · 4.9GB · needs ~5.4GB (4k ctx)
→ whisper:large-v3 on Tesla M60 (16GB, used) ($35) large-v3 · 6.2GB · needs ~8.2GB (4k ctx)
→ mistral-nemo:12b on Tesla M60 (16GB, used) ($35) 12B · 7.1GB · needs ~7.7GB (4k ctx)
→ sdxl on Tesla M60 (16GB, used) ($35) SDXL base · 6.9GB · needs ~8.9GB (4k ctx)
→ llava:13b on Tesla M60 (16GB, used) ($35) 13B · 8.0GB · needs ~11.1GB (4k ctx)
→ gemma3:12b on Tesla M60 (16GB, used) ($35) 12B · 8.1GB · needs ~9.6GB (4k ctx)
→ qwen3:14b on Tesla M60 (16GB, used) ($35) 14B · 9.3GB · needs ~9.9GB (4k ctx)
→ deepseek-r1:14b on Tesla M60 (16GB, used) ($35) distill 14B · 9.0GB · needs ~9.8GB (4k ctx)
→ phi4:14b on Tesla M60 (16GB, used) ($35) 14B · 9.1GB · needs ~9.9GB (4k ctx)
→ gpt-oss:20b on Tesla M40 (24GB, used) ($50) 20B MoE · 14GB · needs ~14.2GB (4k ctx)
→ devstral:24b on Tesla M40 (24GB, used) ($50) 24B · 14.6GB · needs ~15.2GB (4k ctx)
→ qwen3.8:27b on Tesla V100 32GB (used) ($700) 27B · 16.5GB · needs ~18.5GB (4k ctx)
→ gemma3:27b on Tesla V100 32GB (used) ($700) 27B · 17GB · needs ~18.9GB (4k ctx)
→ flux:schnell on Tesla V100 32GB (used) ($700) schnell fp8 · 17GB · needs ~19GB (4k ctx)
→ qwen3:30b-a3b on Tesla V100 32GB (used) ($700) 30B-A3B MoE · 18.6GB · needs ~19GB (4k ctx)
→ qwen3-coder:30b-a3b on Tesla V100 32GB (used) ($700) 30B-A3B MoE · 18.6GB · needs ~19GB (4k ctx)
→ qwen3:32b on Tesla V100 32GB (used) ($700) 32B · 20GB · needs ~21GB (4k ctx)
→ deepseek-r1:32b on Tesla V100 32GB (used) ($700) distill 32B · 20GB · needs ~21GB (4k ctx)
→ qwen2.5vl:32b 32B — 23.0 of 22 GB usable — barely over; real with q8 KV cache (halves KV) or shorter context
→ mixtral:8x7b 8x7B MoE — 26.5 GB vs 22 GB usable — partial CPU offload, expect large speed loss
→ flux:dev dev fp8 — 25.0 GB vs 22 GB usable — partial CPU offload, expect large speed loss
⚠ No measured RTX Spark decode numbers exist yet - the computed band uses the same deliberately wide provisional window as the GB10 card (0.25-0.65); tighten after independent llama.cpp rows land
⚠ NVIDIA's RTX Spark materials quote up to 300 GB/s; the shared GB10-class silicon is documented at 273 GB/s (LPDDR5X, 256-bit) and this card computes with 273
⚠ 45-80W laptop and 140W desktop variants share the silicon; decode is bandwidth-bound so the band does not split by power - prefill does
⚠ Microsoft claims: models over 120B parameters on-device, up to 1 petaflop of AI compute = theoretical FP4 with the sparsity feature per Microsoft's own footnote - vendor claims, not measurements
$5000 — much higher memory bandwidth for models that fit inside 32GB
$16499 — 96GB discrete Blackwell when decode speed matters more than unified-memory capacity