Best home rig for Nemotron 3.5 Bolt 74B-A9B

Will it fit my GPU? Best value GPU per euro →

Real, measured generation speeds for Nemotron 3.5 Bolt 74B-A9B on home hardware, fastest first. Different quants/engines aren't directly comparable.

Build	Quant	Engine	tok/s
RTX 4090: 2026 models (llamabench community)	llama-bench tg128 (quant per result)	llama.cpp	108	source
DGX Spark (GB10, 128GB): Nemotron 3.5 Super VL 120B and Bolt 74B, launch-week runs	2-bit 24GB GGUF + DFlash speculative decoding (7 draft tokens)	llama.cpp	61.61	source
DGX Spark (GB10, 128GB): Nemotron 3.5 Super VL 120B and Bolt 74B, launch-week runs	2-bit 24GB GGUF, no speculative decoding	llama.cpp	39.47	source

Rent before you buy: try this model on a cloud GPU for a few dollars first: Runpod (referral link; sign up with Google to get the new-user bonus)

Memory math: how much VRAM / RAM does Nemotron 3.5 Bolt 74B-A9B need?

Nemotron 3.5 Bolt 74B-A9B has 74.04B parameters, about 9B active per token (source: via llamabench.ai). Weight size = parameters × bits-per-weight ÷ 8, before context (KV cache) and runtime overhead.

Format	Bits/weight	Weights	Smallest common memory size that holds the weights alone
Q8_0	8.50	78.7 GB	96 GB
Q6_K	6.56	60.7 GB	64 GB
Q5_K_M	5.70	52.8 GB	64 GB
Q4_K_M	4.89	45.3 GB	48 GB
Q3_K_M	4.00	37.0 GB	48 GB
Q2_K	3.16	29.2 GB	32 GB

Bits-per-weight for GGUF quants are llama.cpp's own figures (llama.cpp quantize README). "As measured" rows apply the source's stated bits-per-weight to the parameter count; real files differ slightly because some layers stay at higher precision. Try other combinations in the VRAM calculator.

Why it can still run on smaller GPUs: it is a mixture-of-experts model, so only about 9B of its 74.04B parameters are used for each generated token. Runtimes such as llama.cpp (--n-cpu-moe), ExLlamaV3 or Project Maya can keep most expert weights in system RAM and only the shared layers and hot experts in VRAM. Then you need VRAM + system RAM ≥ weights, and speed is limited mainly by RAM bandwidth.

Measured on setups with less GPU memory than the Q4_K_M weights (so part of the model lived in system RAM or was streamed from disk):

RTX 4090 (24 GB): 108 tok/s (llama-bench tg128 (quant per result))

FAQ

How much VRAM does Nemotron 3.5 Bolt 74B-A9B need?

The weights are about 45.3 GB at Q4_K_M (74.04B params × 4.89 bits ÷ 8, before context/KV cache). Add memory for context (KV cache) and the runtime. The smallest-memory setup with a measured result here, fully in memory, is DGX Spark (GB10) (128 GB) at 61.61 tok/s.

What is the fastest measured hardware for Nemotron 3.5 Bolt 74B-A9B?

RTX 4090: 108 tok/s generation (llama.cpp, llama-bench tg128 (quant per result)), from https://llamabench.ai/browse/rtx4090.

Which GPUs have measured results for Nemotron 3.5 Bolt 74B-A9B?

DGX Spark (GB10), RTX 4090.

Can it run on… DGX Spark (GB10) · RTX 4090