Nemotron 3.5 Lightning 30B-A3B on Radeon AI PRO R9700 workstation (32 GB)
31.578B (3B active) on 32 GB at 644 GB/s. Hardware details · Model details
Yes. Nemotron 3.5 Lightning 30B-A3B fits on Radeon AI PRO R9700 workstation (32 GB) at the recommended
Q5_K_M configuration, requiring approximately 23.3 GB at 8K context. Practical context capacity is 512K. Expected decode for the recommended configuration is ~167–241 t/s. EstimatedCompatibilityComfortable
Calculated fit
Yes
Calculated Uses under 80% of usable memory. Room for a long context and other work.
Recommended quantization
Q5_K_M
Calculated gguf
Memory required
23.3 GB
Calculated of 29.4 GB usable — 79%
Max practical context
512K
Calculated model supports 1M
Memory budget at 8K context
Model weights ⓘ
21.5 GB Calculated
KV cache ⓘ
0.1 GB Calculated
Runtime overhead ⓘ
1.6 GB Estimated
Total required
23.3 GB Calculated
Headroom ⓘ
6.2 GB Calculated
Highest-precision quantization that leaves headroom: uses 79% of usable memory at 8K context.
Discrete GPU: 32 GB of VRAM, of which we assume 92% is usable after driver and context overhead.
Mixture of experts: all 31.578B parameters must be resident in memory even though only ~3B are active per token. Memory follows total parameters; speed follows active parameters.
Hybrid cache: token-growing KV memory is charged only to 6 attention blocks; 23 Mamba-2 blocks use fixed-size convolution and recurrent state instead.
Performance
Estimated performance · recommended Q5_K_M
Decode, prefill and TTFT below are estimates for Q5_K_M. We do not have a comparable Q5_K_M measurement on this machine.
Estimated decode
~167–241 t/s
Estimated Q5_K_M; calculated range
Estimated prefill
~2530–5240 t/s
Estimated Q5_K_M; calculated range
Estimated TTFT at 8K
~1.6–3.4 s
Estimated Q5_K_M; calculated range
Hardware load reference
420 W
Measured machine-level load; not this model run
How the estimate is calculated
- Decode: reading 3.0B active parameters at 5.69 bits/weight takes 3.49 ms at 644 GB/s x 95% achieved efficiency.
- MoE routing penalty of 15% applied: expert gathers are less bandwidth-efficient than a dense sweep.
- Prefill: 96 TFLOPS (FP16) x 2.46 calibrated against measured prefill on this platform x 32% assumed model-FLOPs utilisation, divided by 2 x 9.7B parameters per token.
- Prefill uses 9.7B effective parameters, not the 3B active in decode: a batch of hundreds of tokens routes across most of the expert pool.
- MoE decode depends on how well the engine batches expert gathers; real results vary more than for dense models.
- Prefill throughput is highly engine-dependent. Flash attention, batch size and quantized KV all move this number substantially, and for sparse mixture-of-experts models it is the least reliable figure we produce.
- This is a calculated estimate, not a measurement. It assumes a single request, a short prompt, no speculative decoding and a warm model already resident in memory.
Estimator version estimator_v1. Stored with every estimated row so old estimates can be regenerated when the model improves.
Submit a benchmarkSubmissions stay pending until reviewed.
This model at other quantizations on this machine
| Quantization | Resident | Disk | Total needed | Utilisation | Fit | Max context |
|---|---|---|---|---|---|---|
Q4_K_M | 18.4 GB | 18.4 GB | 20.0 GB | 68% | Comfortable | 1M |
Q5_K_Mrecommended | 21.5 GB | 21.5 GB | 23.3 GB | 79% | Comfortable | 512K |
Q8_0 | 31.9 GB | 31.9 GB | 33.9 GB | 115% | Does not fit | 0 |
BF16 | 58.8 GB | 58.8 GB | 61.7 GB | 209% | Does not fit | 0 |
Cost of running Nemotron 3.5 Lightning 30B-A3B on this machineUnited States (federal) · C corporation · 8h/day
Monthly economic cost
$34
Calculated after tax
Codex/Claude Code
$100/mo
≈ $100/mo · local is $66 less
Net cash at purchase
$2,521
Calculated VAT not reclaimable
Total over 5 years
$2,044
Calculated after tax, after resale
Cost per USD/1M tokens
$1.32
Calculated 25.8M tokens/month
The $100 comparison uses the Codex Pro 5x / Claude Max 5x plans. A fixed planning conversion is used for USD. This compares monthly spend only: subscriptions have usage limits, local hardware has different capabilities and constraints, and taxes or regional pricing may change the charged amount. Prices checked 7 September 2026.
Purchase-price input: $2,521 in United States (federal), tax/VAT excluded · estimated · NL retail (system build) (converted catalogue estimate). This is the localized purchase input; the after-tax economic cost is calculated separately below.
Monthly breakdown
| Depreciation 2,143 over 5 years, straight-line to a 378 residual | $35.72 |
| Electricity 21.1 kWh/month at 0.140/kWh | $2.96 |
| Cost of capital 4.0%/yr on 1,450 average capital employed | $4.83 |
| Monthly cost before tax | $43.51 |
| Electricity tax shield Running costs are deductible business expenses | −$0.62 |
| First-year expensing §179 (100.0%) | −$8.82 |
| Monthly economic cost after tax | $34.06 |
Three different numbers, deliberately
Cash cost
$2,521
Money that leaves the bank account on day one, net of reclaimable VAT.
Accounting depreciation
$35.72/month
$2,143 written down over 5 years to a $378 residual.
After-tax economic cost
$34.06/month
Depreciation plus running costs plus cost of capital, less the tax those deductions save. This is the figure to compare between machines.
Assumptions and sources (verified 2026-09-07)
VAT rate
0% Official spec
VAT recoverable
0% Official spec
Federal C-corporation estimate at the flat 21% rate. Pass-through entities should use the sole-proprietor estimate as a rougher proxy.
Effective deduction rate
21.00% Calculated
Headline marginal rate 21.00%.
Depreciation
5 years, straight-line Assumption
Residual value
$378 (15%) Assumption
Two GPU generations later, the card is worth a fraction of its list price.
Electricity
$0.140/kWh Assumption
Average power draw
120 W Calculated
Load 420 W for 20% of powered hours, idle 45 W for the rest — a machine that is on is not generating tokens the whole time.
Investment allowances
§179 100.0% Official spec
Worth $529 in total — first-year expensing that replaces later tax depreciation.
Cost of capital
$4.83/month Assumption
- VAT rate: No US federal VAT · verified 2026-09-07
- Marginal tax rate: IRS Publication 542 — Corporations · verified 2026-09-07
- VAT recoverable fraction: site assumption · verified 2026-09-07
- Useful life: IRS Publication 946 — How To Depreciate Property · verified 2026-09-07
- Electricity price: Site assumption · verified 2026-09-07
Assumes the machine generates tokens 20% of its 176 powered hours per month. Cost per token scales inversely with this number — halve the utilisation and the cost per token doubles.
Federal planning estimate only, not tax advice. State and local income tax, sales/use tax and incentives are excluded. It assumes 100% business use, a Section 179 election, enough business income to use it, and that the full annual limit remains available.
Efficiency metrics
Decode per $1,000 spent
80.9 t/s
Calculated purchase price only
Decode per $100/month
598.7 t/s
Calculated after-tax ownership cost
Tokens per joule
1.70
Calculated same as tokens/s per watt
USD per 1M output tokens
$1.32
Calculated at 20% utilisation
Local versus hosted APIs
| Hosted model | USD/1M output | Break-even | API at your volume | Verdict | Comparison type |
|---|---|---|---|---|---|
| Qwen3.8 Flash Alibaba Cloud | $0.42 | 22.1M/mo | $40 | Local cheaper | Different model |
| DeepSeek-V4.1 Flash DeepSeek | $0.60 | 18.9M/mo | $47 | Local cheaper | Different model |
| DeepSeek-V4 Pro DeepSeek | $1.98 | 4.7M/mo | $188 | Local cheaper | Different model |
| DeepSeek V4 Pro (Together) Together AI | $4.40 | 1.6M/mo | $548 | Local cheaper | Different model |
| GLM-5.3 Z.ai | $4.40 | 2.2M/mo | $403 | Local cheaper | Different model |
| Claude Haiku 4.5 Anthropic | $5.00 | 2.6M/mo | $336 | Local cheaper | Different model |
Different-model comparison: the hosted model is not the model you would run locally. Treat this as a workload-quality trade-off, not a direct economic equivalence — the frontier model may complete a task in fewer tokens, or complete tasks the local model cannot. Break-even is the monthly output volume at which API spend equals the $34/month economic cost of owning this machine, assuming 8 input tokens per output token. Hosted prices are published in USD. This table is the canonical US default.
Nemotron 3.5 Lightning 30B-A3B on other hardware
| Hardware | Decode | Price | Fit |
|---|---|---|---|
| NVIDIA DGX H200 (8x H200, 1,128 GB) | ~533–768 t/s | $359,430 * | Comfortable |
| RTX 5090 workstation (1x 32 GB) | ~319–459 t/s | $3,332 * | Comfortable |
| Quad RTX 5090 workstation (4x 32 GB) | ~315–454 t/s | $12,611 * | Comfortable |
| Lenovo ThinkStation PX (4x RTX PRO 6000, 384 GB) | ~293–422 t/s | $58,553 * | Comfortable |
| Dual RTX 5090 workstation (2x 32 GB) | ~249–359 t/s | $5,854 * | Comfortable |
| Used RTX 4090 workstation (24 GB) | ~237–342 t/s | $2,071 * | Borderline |
Nearest alternatives to the Radeon AI PRO R9700 workstation (32 GB)
Nemotron 3.5 Lightning 30B-A3B on RTX 5090 workstation (1x 32 GB)32 GBNemotron 3.5 Lightning 30B-A3B on Used RTX 4090 workstation (24 GB)24 GBNemotron 3.5 Lightning 30B-A3B on RTX 3090 workstation (Ryzen 9 7950X, 64 GB)24 GBNemotron 3.5 Lightning 30B-A3B on Dual used RTX 4090 workstation (48 GB)48 GBNemotron 3.5 Lightning 30B-A3B on Mac Studio M5 Max 36 GB36 GB
Compare Radeon AI PRO R9700 workstation (32 GB) against RTX 5090 workstation (1x 32 GB) →
Compare Radeon AI PRO R9700 workstation (32 GB) against RTX 5090 workstation (1x 32 GB) →