Cheapest box to run Gemma-3n-E4B at Q4_K_M
Google’s Gemma-3n-E4B is 6.8B 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-3n-E4B at Q4_K_M needs 6GB.
Cheapest Config that fits: Mac mini, M6, 16GB. Fits, 7GB headroom. Est. 45.8 tok/s.
Best offer today: $1,099 at Apple — in stock, checked 35 hours ago.
Collecting price history — 3 of 90 days. No timing verdict yet.
Google names this model for its 4B effective parameters, but 6.8B raw weights are what you load, so the memory arithmetic uses 6.8B. The K/V figure accounts for layer sharing and sliding-window attention and is an estimate.
Offers — Mac mini, M6, 16GB, 512GB
- Applelist pricein stockchecked 35 hours ago$1,099
- Mac mini, M6, 16GB$1,099Fits7GB headroomest. 45.8 tok/s
- Mac mini, M6, 24GB$1,299Fits15GB headroomest. 45.8 tok/s
- Mac mini, M6, 32GB$1,499Fits23GB headroomest. 45.8 tok/s
- Mac mini, M5 Pro, 24GB$1,699Fits15GB headroomest. 82.6 tok/s
- Mac mini, M5 Pro, 48GB$2,499Fits38GB headroomest. 82.6 tok/s
- Mac Studio, M5 Max, 36GB$2,499Fits27GB headroomest. 165 tok/s
- Mac mini, M5 Pro, 64GB$2,899Fits53GB headroomest. 82.6 tok/s
- GeForce RTX 5090, 32GB$2,999.99+hostFits25GB headroomest. 482 tok/s
- Fits53GB headroomest. 147 tok/s
- Mac Studio, M5 Max, 48GB$3,099Fits38GB headroomest. 165 tok/s
- Mac Studio, M5 Max, 64GB$3,499Fits53GB headroomest. 165 tok/s
- Fits114GB headroomest. 73.5 tok/s
- Fits82GB headroomest. 220 tok/s
- Fits114GB headroomest. 147 tok/s
- Fits114GB headroomest. 73.5 tok/s
- Fits114GB headroomest. 165 tok/s
- Fits82GB headroomest. 323 tok/s
- Fits242GB headroomest. 220 tok/s
- Fits242GB headroomest. 323 tok/s
- Fits498GB headroomest. 323 tok/s
| Config | Memory | Price | $/GB | Fit | Headroom | Est. speed |
|---|---|---|---|---|---|---|
| Mac mini, M6 | 16GB | $1,099 | $69 | Fits | 7GB | 45.8 tok/s |
| Mac mini, M6 | 24GB | $1,299 | $54 | Fits | 15GB | 45.8 tok/s |
| Mac mini, M6 | 32GB | $1,499 | $47 | Fits | 23GB | 45.8 tok/s |
| Mac mini, M5 Pro | 24GB | $1,699 | $71 | Fits | 15GB | 82.6 tok/s |
| Mac mini, M5 Pro | 48GB | $2,499 | $52 | Fits | 38GB | 82.6 tok/s |
| Mac Studio, M5 Max | 36GB | $2,499 | $69 | Fits | 27GB | 165 tok/s |
| Mac mini, M5 Pro | 64GB | $2,899 | $45 | Fits | 53GB | 82.6 tok/s |
| GeForce RTX 5090 | 32GB | $2,999.99+host | $94 | Fits | 25GB | 482 tok/s |
| Mac Studio, M4 Max, refurb | 64GB | $3,049 | $48 | Fits | 53GB | 147 tok/s |
| Mac Studio, M5 Max | 48GB | $3,099 | $65 | Fits | 38GB | 165 tok/s |
| Mac Studio, M5 Max | 64GB | $3,499 | $55 | Fits | 53GB | 165 tok/s |
| Ascent GX10, GB10 Grace Blackwell | 128GB | $3,999.99 | $31 | Fits | 114GB | 73.5 tok/s |
| Mac Studio, M3 Ultra, refurb | 96GB | $4,329 | $45 | Fits | 82GB | 220 tok/s |
| Mac Studio, M4 Max, refurb | 128GB | $4,409 | $34 | Fits | 114GB | 147 tok/s |
| DGX Spark, GB10 Grace Blackwell | 128GB | $4,699 | $37 | Fits | 114GB | 73.5 tok/s |
| Mac Studio, M5 Max | 128GB | $5,099 | $40 | Fits | 114GB | 165 tok/s |
| Mac Studio, M5 Ultra | 96GB | $5,499 | $57 | Fits | 82GB | 323 tok/s |
| Mac Studio, M3 Ultra, refurb | 256GB | $7,729 | $30 | Fits | 242GB | 220 tok/s |
| Mac Studio, M5 Ultra | 256GB | $9,499 | $37 | Fits | 242GB | 323 tok/s |
| Mac Studio, M5 Ultra | 512GB | — | — | Fits | 498GB | 323 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 = 6.8B × 0.604 B/param (4.83 bits) + 0.072 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.65 ÷ (4B active × 0.604 B/param) = 2.4 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 16GB runs
- gpt-oss-20bQ3_K_M, est. 48.3 tok/s
- Phi-4Q5_K_M, est. 11.1 tok/s
- Qwen3-VL-8B-InstructQ8_0, est. 13.0 tok/s
- Gemma-3n-E4BQ8_0, est. 26.0 tok/s
- Qwen3-4BBF16, est. 13.8 tok/s
Best quant that still fits at 8K context, largest model first. Full list on the Config page.
Bandwidth 170GB/s, usable memory 13GB of 16GB.