Cheapest box to run Kimi-K2-Instruct at Q8_0
Moonshot’s Kimi-K2-Instruct is 1000B parameters, 32B of them read per token. The number below is computed from that and each vendor’s published specs, not asserted. Switch the quant or the context and it recomputes in place.
Kimi-K2-Instruct at Q8_0 needs 1064GB.
Nothing on the Grid runs Kimi-K2-Instruct at Q8_0 today. The largest Config, Mac Studio, M5 Ultra, 512GB, is 560GB short. At it fits with 13GB headroom.
A trillion parameters. No Config we track holds it at any quant, and we list it so the arithmetic says so plainly.
- Mac mini, M6, 16GB$1,099Won't fit−1051GB headroom
- Mac mini, M6, 24GB$1,299Won't fit−1043GB headroom
- Mac mini, M6, 32GB$1,499Won't fit−1035GB headroom
- Mac mini, M5 Pro, 24GB$1,699Won't fit−1043GB headroom
- Mac mini, M5 Pro, 48GB$2,499Won't fit−1020GB headroom
- Mac Studio, M5 Max, 36GB$2,499Won't fit−1031GB headroom
- Mac mini, M5 Pro, 64GB$2,899Won't fit−1005GB headroom
- GeForce RTX 5090, 32GB$2,999.99+hostWon't fit−1034GB headroom
- Won't fit−1005GB headroom
- Mac Studio, M5 Max, 48GB$3,099Won't fit−1020GB headroom
- Mac Studio, M5 Max, 64GB$3,499Won't fit−1005GB headroom
- Won't fit−944GB headroom
- Won't fit−976GB headroom
- Won't fit−944GB headroom
- Won't fit−944GB headroom
- Won't fit−944GB headroom
- Won't fit−976GB headroom
- Won't fit−816GB headroom
- Won't fit−816GB headroom
- Won't fit−560GB headroom
| Config | Memory | Price | $/GB | Fit | Headroom | Est. speed |
|---|---|---|---|---|---|---|
| Mac mini, M6 | 16GB | $1,099 | $69 | Won't fit | −1051GB | — |
| Mac mini, M6 | 24GB | $1,299 | $54 | Won't fit | −1043GB | — |
| Mac mini, M6 | 32GB | $1,499 | $47 | Won't fit | −1035GB | — |
| Mac mini, M5 Pro | 24GB | $1,699 | $71 | Won't fit | −1043GB | — |
| Mac mini, M5 Pro | 48GB | $2,499 | $52 | Won't fit | −1020GB | — |
| Mac Studio, M5 Max | 36GB | $2,499 | $69 | Won't fit | −1031GB | — |
| Mac mini, M5 Pro | 64GB | $2,899 | $45 | Won't fit | −1005GB | — |
| GeForce RTX 5090 | 32GB | $2,999.99+host | $94 | Won't fit | −1034GB | — |
| Mac Studio, M4 Max, refurb | 64GB | $3,049 | $48 | Won't fit | −1005GB | — |
| Mac Studio, M5 Max | 48GB | $3,099 | $65 | Won't fit | −1020GB | — |
| Mac Studio, M5 Max | 64GB | $3,499 | $55 | Won't fit | −1005GB | — |
| Ascent GX10, GB10 Grace Blackwell | 128GB | $3,999.99 | $31 | Won't fit | −944GB | — |
| Mac Studio, M3 Ultra, refurb | 96GB | $4,329 | $45 | Won't fit | −976GB | — |
| Mac Studio, M4 Max, refurb | 128GB | $4,409 | $34 | Won't fit | −944GB | — |
| DGX Spark, GB10 Grace Blackwell | 128GB | $4,699 | $37 | Won't fit | −944GB | — |
| Mac Studio, M5 Max | 128GB | $5,099 | $40 | Won't fit | −944GB | — |
| Mac Studio, M5 Ultra | 96GB | $5,499 | $57 | Won't fit | −976GB | — |
| Mac Studio, M3 Ultra, refurb | 256GB | $7,729 | $30 | Won't fit | −816GB | — |
| Mac Studio, M5 Ultra | 256GB | $9,499 | $37 | Won't fit | −816GB | — |
| Mac Studio, M5 Ultra | 512GB | — | — | Won't fit | −560GB | — |
GeForce RTX 5090 is a card, not a computer: every price shown for it is the cheapest listing we found plus $1,000 for the minimum host machine.
memory = 1000B × 1.063 B/param (8.5 bits) + 0.070 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.5 ÷ (32B active × 1.063 B/param) = 34.0 GB read per token
usable = memory − operating system reserve (8% on a Mac, 6% on a DGX Spark, 4% on a card)
Bandwidth is the vendor’s published figure. Every speed is an estimate and says so. Methodology.