Cheapest box to run Gemma-3-27B at Q3_K_M
Google’s Gemma-3-27B is 27B parameters, dense. 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.
Gemma-3-27B at Q3_K_M needs 16GB.
Cheapest Config that fits: Mac mini, M6, 24GB. Fits, 5GB headroom. Est. 8.4 tok/s.
Best offer today: $1,299 at Apple — in stock, checked 19 minutes ago.
Price history: 47 days recorded.
Offers — Mac mini, M6, 24GB, 512GB
- Mac mini, M6, 16GB$1,099Won't fit−3GB headroom
- Mac mini, M6, 24GB$1,299Fits5GB headroomest. 8.4 tok/s
- Mac mini, M6, 32GB$1,499Fits13GB headroomest. 8.4 tok/s
- Mac mini, M5 Pro, 24GB$1,699Fits5GB headroomest. 15.1 tok/s
- Mac mini, M5 Pro, 48GB$2,499Fits28GB headroomest. 15.1 tok/s
- Mac Studio, M5 Max, 36GB$2,499Fits17GB headroomest. 22.7 tok/s
- Mac mini, M5 Pro, 64GB$2,899Fits42GB headroomest. 15.1 tok/s
- GeForce RTX 5090, 32GB$2,999.99+hostFits14GB headroomest. 88.3 tok/s
- Fits42GB headroomest. 26.9 tok/s
- Mac Studio, M5 Max, 48GB$3,099Fits28GB headroomest. 30.2 tok/s
- Mac Studio, M5 Max, 64GB$3,499Fits42GB headroomest. 30.2 tok/s
- Fits104GB headroomest. 13.4 tok/s
- Fits72GB headroomest. 40.3 tok/s
- Fits104GB headroomest. 26.9 tok/s
- Fits104GB headroomest. 13.4 tok/s
- Fits104GB headroomest. 30.2 tok/s
- Fits72GB headroomest. 59.1 tok/s
- Fits232GB headroomest. 40.3 tok/s
- Fits232GB headroomest. 59.1 tok/s
- Fits488GB headroomest. 59.1 tok/s
| Config | Memory | Price | Fit | Headroom | Est. speed |
|---|---|---|---|---|---|
| Mac mini, M6 | 16GB | $1,099 | Won't fit | −3GB | — |
| Mac mini, M6 | 24GB | $1,299 | Fits | 5GB | 8.4 tok/s |
| Mac mini, M6 | 32GB | $1,499 | Fits | 13GB | 8.4 tok/s |
| Mac mini, M5 Pro | 24GB | $1,699 | Fits | 5GB | 15.1 tok/s |
| Mac mini, M5 Pro | 48GB | $2,499 | Fits | 28GB | 15.1 tok/s |
| Mac Studio, M5 Max | 36GB | $2,499 | Fits | 17GB | 22.7 tok/s |
| Mac mini, M5 Pro | 64GB | $2,899 | Fits | 42GB | 15.1 tok/s |
| GeForce RTX 5090 | 32GB | $2,999.99+host | Fits | 14GB | 88.3 tok/s |
| Mac Studio, M4 Max, refurb | 64GB | $3,049 | Fits | 42GB | 26.9 tok/s |
| Mac Studio, M5 Max | 48GB | $3,099 | Fits | 28GB | 30.2 tok/s |
| Mac Studio, M5 Max | 64GB | $3,499 | Fits | 42GB | 30.2 tok/s |
| Ascent GX10, GB10 Grace Blackwell | 128GB | $3,999.99 | Fits | 104GB | 13.4 tok/s |
| Mac Studio, M3 Ultra, refurb | 96GB | $4,329 | Fits | 72GB | 40.3 tok/s |
| Mac Studio, M4 Max, refurb | 128GB | $4,409 | Fits | 104GB | 26.9 tok/s |
| DGX Spark, GB10 Grace Blackwell | 128GB | $4,699 | Fits | 104GB | 13.4 tok/s |
| Mac Studio, M5 Max | 128GB | $5,099 | Fits | 104GB | 30.2 tok/s |
| Mac Studio, M5 Ultra | 96GB | $5,499 | Fits | 72GB | 59.1 tok/s |
| Mac Studio, M3 Ultra, refurb | 256GB | $7,729 | Fits | 232GB | 40.3 tok/s |
| Mac Studio, M5 Ultra | 256GB | $9,499 | Fits | 232GB | 59.1 tok/s |
| Mac Studio, M5 Ultra | 512GB | — | Fits | 488GB | 59.1 tok/s |
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 = 27B × 0.489 B/param (3.91 bits) + 0.254 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.65 ÷ (27B active × 0.489 B/param) = 13.2 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.
What else Mac mini M6 24GB runs
- Qwen3.5-35B-A3BQ3_K_M, est. 58.0 tok/s
- Qwen3-32BQ3_K_M, est. 7.1 tok/s
- Gemma-4-31BQ3_K_M, est. 7.4 tok/s
- Qwen3-30B-A3BQ4_K_M, est. 42.7 tok/s
- Gemma-3-27BQ4_K_M, est. 6.8 tok/s
- Qwen3.6-27BQ5_K_M, est. 5.8 tok/s
Best quant that still fits at 8K context, largest model first. Full list on the Config page.
Bandwidth 170GB/s, usable memory 21GB of 24GB.