Sunk Cost sunkcost.ai Data checked 2026-09-03

What hardware do you need to run Ministral 3 14B?

Ministral 3 14B at Q4_K_M is 8.2 GB of weights, with a context ceiling of 256k tokens. Not yet rated: released after our last ratings pass. Matches the older Mistral Small 24B on the index at a third of the size.

Cheapest machine that runs itMac mini M6, 24GB at $1,099
Fastest of the ones listedMac Studio M5 Ultra, 96GB — 66 tok/s at 32k context
Honest answer on costPays back in 34 years at 500k tokens a day.

How good is it, really?

On the Artificial Analysis Intelligence Index v4.3 it scores 6, which puts it in the Below every hosted tier band. Fine for simple, well-specified tasks. Noticeably less capable than anything the big labs sell today. Score source. See the whole table.

What it costs either way

Renting the same model costs $0.2 per million input tokens and $0.2 per million output (OpenRouter, cheapest active endpoint, checked 2026-09-15). Buying a machine only beats that if you use it hard enough, for long enough, that the hardware price divides down below the rental bill.

Machines that run it

MachinePriceSpeed at 32kPay-back
Mac mini M6, 24GB $1,099 9.4 tok/s estimated Pays back in 34 years Run the numbers
Strix Halo Framework Desktop, 32GB $1,269 14 tok/s estimated Pays back in 40 years Run the numbers
MacBook Pro M5 (14-inch), 32GB $2,399 8.4 tok/s estimated Pays back in 74 years Run the numbers
Mac Studio M5 Max, 36GB $2,499 25 tok/s estimated Pays back in 75 years Run the numbers
DGX Spark GB10 Grace Blackwell, 128GB $4,699 15 tok/s estimated Pays back in 151 years Run the numbers

One machine per family, cheapest first. Speeds are measured where a public benchmark exists and estimated from memory bandwidth otherwise; the calculator says which for any configuration.

The specifics

Parameters
14B
Quantisation
Q4_K_M
Weights on disk
8.2 GB
KV cache
5.4 GB at 32k context — All 40 layers full attention; no sliding window.
Maximum context
256k tokens (256k)
Licence
Apache 2.0
Sources
source 1