modelrigs

Phi-3.5 MoE 42B-A6.6B on NVIDIA RTX A5000 24 GB

41.873B (6.6B active) on 24 GB at 768 GB/s. Hardware details · Model details

No calculated fit. Phi-3.5 MoE 42B-A6.6B does not fit under standard overhead assumptions at Q4_K_M: the model calculates 27.1 GB required against 22.1 GB usable.
CompatibilityDoes not fit
Calculated fit
No
Calculated Requires more than 97% of usable memory. Needs a smaller quantization or more memory.
Smallest catalogued quantization
Q4_K_M
Calculated gguf
Memory required
27.1 GB
Calculated of 22.1 GB usable — 123%
Max practical context
0
Calculated model supports 128K

Memory budget at 8K context

Model weights ⓘ
24.4 GB Calculated
KV cache ⓘ
1.0 GB Calculated
Runtime overhead ⓘ
1.7 GB Estimated
Total required
27.1 GB Calculated
Headroom ⓘ
-5.0 GB Calculated
No quantization in our catalogue fits this machine.
Discrete GPU: 24 GB of VRAM, of which we assume 92% is usable after driver and context overhead.
Mixture of experts: all 41.873B parameters must be resident in memory even though only ~6.6B are active per token. Memory follows total parameters; speed follows active parameters.
This model can be run with layers offloaded to system RAM, but decode throughput typically drops by an order of magnitude once any significant fraction of the weights crosses PCIe. We do not count offload as fitting.
Performance

Estimated performance · recommended Q4_K_M

Decode, prefill and TTFT below are estimates for Q4_K_M. We do not have a comparable Q4_K_M measurement on this machine.

Estimated decode
—
Estimated Q4_K_M; calculated range
Estimated prefill
—
Estimated Q4_K_M; calculated range
Estimated TTFT at 8K
—
Estimated Q4_K_M; calculated range
Hardware load reference
230 W
Official spec hardware TDP; not this model run
No throughput figures for this configuration. We do not estimate performance for a model that cannot load.
How the estimate is calculated
  1. Decode: reading 6.6B active parameters at 4.85 bits/weight takes 7.44 ms at 768 GB/s x 70% achieved efficiency.
  2. MoE routing penalty of 15% applied: expert gathers are less bandwidth-efficient than a dense sweep.
  • MoE decode depends on how well the engine batches expert gathers; real results vary more than for dense models.
  • No FP16 compute figure for this hardware, so prefill cannot be estimated.
  • 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
QuantizationResidentDiskTotal neededUtilisationFitMax context
Q4_K_Msmallest listed24.4 GB24.4 GB27.1 GB123%Does not fit0
Q5_K_M28.6 GB28.6 GB31.4 GB142%Does not fit0
Q8_042.3 GB42.3 GB45.5 GB206%Does not fit0
BF1678.0 GB78.0 GB82.3 GB373%Does not fit0
Cost of running Phi-3.5 MoE 42B-A6.6B on this machineUnited States (federal) · C corporation · 8h/day
Monthly economic cost
$1
Calculated after tax
Codex/Claude Code
$100/mo
≈ $100/mo · local is $99 less
Net cash at purchase
$0
Calculated VAT not reclaimable
Total over 5 years
$80
Calculated after tax, after resale
Cost per USD/1M tokens
—
Calculated

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.

Monthly breakdown

Depreciation
0 over 5 years, straight-line to a 0 residual
$0.00
Electricity
12.0 kWh/month at 0.140/kWh
$1.68
Monthly cost before tax$1.68
Electricity tax shield
Running costs are deductible business expenses
−$0.35
Monthly economic cost after tax$1.33

Three different numbers, deliberately

Cash cost
$0
Money that leaves the bank account on day one, net of reclaimable VAT.
Accounting depreciation
$0.00/month
$0 written down over 5 years to a $0 residual.
After-tax economic cost
$1.33/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
$0 (15%) Assumption
Two GPU generations later, the card is worth a fraction of its list price.
Electricity
$0.140/kWh Assumption
Average power draw
68 W Calculated
Load 230 W for 20% of powered hours, idle 28 W for the rest — a machine that is on is not generating tokens the whole time.
Investment allowances
§179 100.0% Official spec
Worth $0 in total — first-year expensing that replaces later tax depreciation.
Cost of capital
not applied Assumption
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.
Phi-3.5 MoE 42B-A6.6B on other hardware
HardwareDecodePriceFit
Nearest alternatives to the NVIDIA RTX A5000 24 GB
Phi-3.5 MoE 42B-A6.6B on GeForce RTX 3090 Ti 24 GB24 GBPhi-3.5 MoE 42B-A6.6B on NVIDIA Tesla P40 24 GB24 GBPhi-3.5 MoE 42B-A6.6B on RTX 3090 workstation (Ryzen 9 7950X, 64 GB)24 GBPhi-3.5 MoE 42B-A6.6B on Used RTX 4090 workstation (24 GB)24 GBPhi-3.5 MoE 42B-A6.6B on NVIDIA RTX 5000 Ada 32 GB32 GB
Compare NVIDIA RTX A5000 24 GB against GeForce RTX 3090 Ti 24 GB →