modelrigs

Phi-3.5 MoE 42B-A6.6B on Mac Studio M3 Ultra 96 GB

41.873B (6.6B active) on 96 GB at 819 GB/s. Hardware details · Model details

Yes. Phi-3.5 MoE 42B-A6.6B fits on Mac Studio M3 Ultra 96 GB at the recommended Q8_0 configuration, requiring approximately 45.5 GB at 8K context. Practical context capacity is 128K.
CompatibilityComfortable
Calculated fit
Yes
Calculated Uses under 80% of usable memory. Room for a long context and other work.
Recommended quantization
Q8_0
Calculated gguf
Memory required
45.5 GB
Calculated of 81.6 GB usable — 56%
Max practical context
128K
Calculated model supports 128K

Memory budget at 8K context

Model weights ⓘ
42.3 GB Calculated
KV cache ⓘ
1.0 GB Calculated
Runtime overhead ⓘ
2.3 GB Estimated
Total required
45.5 GB Calculated
Headroom ⓘ
36.1 GB Calculated
Highest-precision quantization that leaves headroom: uses 56% of usable memory at 8K context.
Unified memory: weights, KV cache and the OS share the same 96 GB pool. We assume 85% is addressable by the GPU, which matches a headless machine with the wired-memory limit raised.
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.
Performance

Estimated performance · recommended Q8_0

Decode, prefill and TTFT below are estimates for Q8_0. We do not have a comparable Q8_0 measurement on this machine.

Estimated decode
—
Estimated Q8_0; calculated range
Estimated prefill
—
Estimated Q8_0; calculated range
Estimated TTFT at 8K
—
Estimated Q8_0; calculated range
Hardware load reference
—
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 8.50 bits/weight takes 12.09 ms at 819 GB/s x 71% 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
MLX 4-bit22.8 GB22.8 GB25.5 GB31%Comfortable128K
Q4_K_M24.4 GB24.4 GB27.1 GB33%Comfortable128K
Q5_K_M28.6 GB28.6 GB31.4 GB39%Comfortable128K
Q8_0recommended42.3 GB42.3 GB45.5 GB56%Comfortable128K
BF1678.0 GB78.0 GB82.3 GB101%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
$0
Calculated after tax
Codex/Claude Code
$100/mo
≈ $100/mo · local is $100 less
Net cash at purchase
$0
Calculated VAT not reclaimable
Total over 5 years
$0
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
0.0 kWh/month at 0.140/kWh
$0.00
Monthly cost before tax$0.00
Electricity tax shield
Running costs are deductible business expenses
-$0.00
Monthly economic cost after tax$0.00

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
$0.00/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 (30%) Assumption
Extrapolated from the three-year curve.
Electricity
$0.000/kWh Assumption
Average power draw
0 W Calculated
Load — for 20% of powered hours, idle 0 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 Mac Studio M3 Ultra 96 GB
Phi-3.5 MoE 42B-A6.6B on Mac Studio M3 Ultra 256 GB256 GBPhi-3.5 MoE 42B-A6.6B on Mac Studio M3 Ultra 512 GB512 GBPhi-3.5 MoE 42B-A6.6B on Mac Studio M2 Max 96 GB96 GBPhi-3.5 MoE 42B-A6.6B on Mac Studio M5 Ultra 96 GB96 GBPhi-3.5 MoE 42B-A6.6B on Mac Studio M1 Ultra 128 GB128 GB
Compare Mac Studio M3 Ultra 96 GB against Mac Studio M3 Ultra 256 GB →