Can the GeForce RTX 5090 run NVIDIA Nemotron 3.5 Lightning 30B-A3B?
Yes. A new GeForce RTX 5090 ($6,500) runs NVIDIA Nemotron 3.5 Lightning 30B-A3B at Q4 with room for its full 256K context, at about 177 tok/s estimate. At $37 per tok/s it is the cheapest of the 2 priced GPUs that fit it.
Yes, with room for long context. On the NVIDIA GeForce RTX 5090 (32 GB) there is room for about its full 256K tokens of context with the default FP16 KV cacheestimate: what is left after the weights and a 1 GB runtime reserve. The Q4_K_M GGUF of NVIDIA Nemotron 3.5 Lightning 30B-A3B is 25.48 GB (source); with our 1.5 GB allowance for the runtime and a short context it needs about 27 GB.
NVIDIA Nemotron 3.5 Lightning 30B-A3B fits 2 of the 18 GPUs we have a dated price for; the smallest of them has 32 GB. The GeForce RTX 5090's 1792 GB/s of memory bandwidth makes it the fastest of the 2, at ~177 tok/s estimate. Only 3B of its 31.58B parameters are active per token, so the estimate reads 2.42 GB per token rather than the whole 25.48 GB file, plus a per-token routing and expert-gather cost calibrated to 45 measured llama.cpp benchmarks (2025–26). On the same card, the mixture-of-experts Recursive Flint 40B-A2.5B (2.5B of 39.8B active, 25.99 GB) estimates ~366 tok/s.
The cheapest way to run it per tok/s. At $6,500 (new) the GeForce RTX 5090 works out to $37 per tok/s; the next best is the RTX A6000 at $64. The cheapest complete build we priced that runs NVIDIA Nemotron 3.5 Lightning 30B-A3B is our 32 GB: two Radeon RX 9070 build at $3,589.93.
Last updated Oct 16, 2026. Speed is a bandwidth-based estimate (1792 GB/s, NVIDIA), scaled by per-platform constants calibrated to 45 measured llama.cpp benchmarks (2025–26), not a benchmark of this pair. Max context is an estimate from the model's architecture for llama.cpp with flash attention.
Next size up: Ripple-65B Preview (38.8 GB at Q4_K_M) needs ~40.3 GB and does not fit on one NVIDIA GeForce RTX 5090. See what runs Ripple-65B Preview.
Every priced GPU that runs NVIDIA Nemotron 3.5 Lightning 30B-A3B, by cost per token/sec
| GPU | VRAM | Speed estimate | Max context estimate | Low price | $ per tok/s |
|---|---|---|---|---|---|
| NVIDIA GeForce RTX 5090 | 32 GB | ~177 tok/s | 256K (full) | $6,500 (new) | $37 |
| NVIDIA RTX A6000 | 48 GB | ~130 tok/s | 256K (full) | $8,363 (new) | $64 |
Prices: picksbycard.com, 2026-11-02; see how prices are sourced. Prices move daily. Bar length shows value: longer is cheaper per tok/s.