Skip to main contentThe Local Bench

Illustrative build · sourced prices

Strix Halo SFF 128 GB

128 GB of unified memory in a mini PC. Holds gpt-oss-120b with room to spare; dense 70B models are slow.

Mini PC
  • Data checked: 2026-10-12. Review interval shown below. Prices need review after 30 days; specifications and benchmark collections after 90 days. Dates describe the recorded check, not a live retailer quote.

    • Verified Checked against the primary source (for example the model’s config.json on Hugging Face).
    • Sourced Taken from a cited source; not yet independently checked.
    • Estimate Calculated with the published formulas; not a measurement.
    • Sample Placeholder data for layout; not a real figure.
    • Not yet measured No public, citable score found for this benchmark yet. We never fill the gap with a guess.

    Verified metadata is not a measured performance result. Only cited runs matching model, device, quant and recorded context can replace speed estimates. Estimates assume one stream and F16 KV cache, with a ±25% uncertainty band. A published run applies only to its recorded runtime and conditions; missing settings are unknown.

    Evidence labels · Freshness policy · Calculation methodology

$3,499.99
complete-system parts, USDSourced
37.5tok/s
gpt-oss-120bEstimate
120W
at load; 15 W idle
17.8months
illustrative break-even against a $200/month plan

Includes the listed system parts. Excludes tax, shipping, peripherals, maintenance, financing and resale. Sample parts prices may differ from current hardware catalogue prices. Electricity assumes 120 W whole-system load for 4 h and 15 W idle for 20 h daily, 30 days at $0.15/kWh. Power is unmeasured.

Parts list

Price from CFG (Oct 2026); 120 W load is the brief's break-even assumption.

Part
Item
Price, USD
SystemRyzen AI Max+ 395 mini PC$3,499.99Sourced

No affiliate links yet. Prices will be timestamped per retailer once tracking starts.

Expected speed

Reference models, default quant, 8k context.

Model
Verdict
tok/s
gpt-oss-120bFits37.5Estimate
Qwen3 30B-A3BFits41.0Estimate
Llama 3.3 70B InstructFits2.5Estimate

Break-even card

Electricity $3.51/month (120 W × 4 h + 15 W idle × 20 h, $0.15/kWh).

Plan
Price / month
Months to break even
ChatGPT Plus$20212.2
Claude Pro$20212.2
Google AI Pro$19.99212.4
Claude Max (5×)$10036.3
ChatGPT Pro 100$10036.3
Google AI Ultra (5×)$99.9936.3
Claude Max (20×)$20017.8
ChatGPT Pro 200$20017.8
Google AI Ultra (20×)$199.9917.8
ChatGPT Pro 500 (top tier)$5007.0

Break-even = hardware ÷ (plan − electricity). It ignores capability: a local model is not a frontier model. Adjust the numbers.