Cheapest box to run GLM-4.6 at Q4_K_M
Z.ai’s GLM-4.6 is 357B 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.
GLM-4.6 at Q4_K_M needs 220GB.
Cheapest Config that fits: Mac Studio, M3 Ultra, 256GB, refurbished. Fits, 28GB headroom. Est. 21.2 tok/s.
Best offer today: $7,729 at Apple Certified Refurbished — in stock, checked 1 hour ago.
Collecting price history — 3 of 90 days. No timing verdict yet.
Offers — Mac Studio, M3 Ultra, 256GB, 1TB, refurbished
- Apple Certified RefurbishedRefurbished by Apple, with Apple's one-year limited warranty and 14-day returns. Apple says refurbished supply is usually very limited and often runs out, so this row can go dead faster than a new one.in stockchecked 1 hour ago$7,729
- Mac mini, M6, 16GB$1,099Won't fit−207GB headroom
- Mac mini, M6, 24GB$1,299Won't fit−199GB headroom
- Mac mini, M6, 32GB$1,499Won't fit−191GB headroom
- Mac mini, M5 Pro, 24GB$1,699Won't fit−199GB headroom
- Mac mini, M5 Pro, 48GB$2,499Won't fit−176GB headroom
- Mac Studio, M5 Max, 36GB$2,499Won't fit−187GB headroom
- Mac mini, M5 Pro, 64GB$2,899Won't fit−161GB headroom
- GeForce RTX 5090, 32GB$2,999.99+hostWon't fit−189GB headroom
- Won't fit−161GB headroom
- Mac Studio, M5 Max, 48GB$3,099Won't fit−176GB headroom
- Mac Studio, M5 Max, 64GB$3,499Won't fit−161GB headroom
- Won't fit−99GB headroom
- Won't fit−131GB headroom
- Won't fit−100GB headroom
- Won't fit−99GB headroom
- Won't fit−100GB headroom
- Won't fit−131GB headroom
- Fits28GB headroomest. 21.2 tok/s
- Fits28GB headroomest. 31.1 tok/s
- Fits284GB headroomest. 31.1 tok/s
| Config | Memory | Price | $/GB | Fit | Headroom | Est. speed |
|---|---|---|---|---|---|---|
| Mac mini, M6 | 16GB | $1,099 | $69 | Won't fit | −207GB | — |
| Mac mini, M6 | 24GB | $1,299 | $54 | Won't fit | −199GB | — |
| Mac mini, M6 | 32GB | $1,499 | $47 | Won't fit | −191GB | — |
| Mac mini, M5 Pro | 24GB | $1,699 | $71 | Won't fit | −199GB | — |
| Mac mini, M5 Pro | 48GB | $2,499 | $52 | Won't fit | −176GB | — |
| Mac Studio, M5 Max | 36GB | $2,499 | $69 | Won't fit | −187GB | — |
| Mac mini, M5 Pro | 64GB | $2,899 | $45 | Won't fit | −161GB | — |
| GeForce RTX 5090 | 32GB | $2,999.99+host | $94 | Won't fit | −189GB | — |
| Mac Studio, M4 Max, refurb | 64GB | $3,049 | $48 | Won't fit | −161GB | — |
| Mac Studio, M5 Max | 48GB | $3,099 | $65 | Won't fit | −176GB | — |
| Mac Studio, M5 Max | 64GB | $3,499 | $55 | Won't fit | −161GB | — |
| Ascent GX10, GB10 Grace Blackwell | 128GB | $3,999.99 | $31 | Won't fit | −99GB | — |
| Mac Studio, M3 Ultra, refurb | 96GB | $4,329 | $45 | Won't fit | −131GB | — |
| Mac Studio, M4 Max, refurb | 128GB | $4,409 | $34 | Won't fit | −100GB | — |
| DGX Spark, GB10 Grace Blackwell | 128GB | $4,699 | $37 | Won't fit | −99GB | — |
| Mac Studio, M5 Max | 128GB | $5,099 | $40 | Won't fit | −100GB | — |
| Mac Studio, M5 Ultra | 96GB | $5,499 | $57 | Won't fit | −131GB | — |
| Mac Studio, M3 Ultra, refurb | 256GB | $7,729 | $30 | Fits | 28GB | 21.2 tok/s |
| Mac Studio, M5 Ultra | 256GB | $9,499 | $37 | Fits | 28GB | 31.1 tok/s |
| Mac Studio, M5 Ultra | 512GB | — | — | Fits | 284GB | 31.1 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 = 357B × 0.604 B/param (4.83 bits) + 0.377 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.5 ÷ (32B active × 0.604 B/param) = 19.3 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 Studio M3 Ultra 256GB refurb runs
- Qwen3-Coder-480B-A35B-InstructQ3_K_M, est. 23.9 tok/s
- Llama-4-Maverick-17B-128EQ4_K_M, est. 39.9 tok/s
- GLM-4.6Q4_K_M, est. 21.2 tok/s
- Qwen3-235B-A22BQ6_K, est. 22.7 tok/s
- MiniMax-M2Q8_0, est. 38.5 tok/s
- Mistral-Large-2411Q8_0, est. 4.1 tok/s
Best quant that still fits at 8K context, largest model first. Full list on the Config page.
Bandwidth 819GB/s, usable memory 248GB of 256GB.