Cheapest box to run Qwen3-Coder-Next at Q8_0
Alibaba’s Qwen3-Coder-Next is 80B parameters, 3B 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.
Qwen3-Coder-Next at Q8_0 needs 86GB.
Cheapest Config that fits: Ascent GX10, GB10 Grace Blackwell, 128GB. Fits, 34GB headroom. Est. 42.8 tok/s.
Best offer today: $3,999.99 at Best Buy — in stock, checked 33 days ago.
Price history: 34 days recorded.
The Qwen3-Next hybrid again: 12 of 48 layers are full attention and the linear layers cache nothing per token.
Offers — Ascent GX10, GB10 Grace Blackwell, 128GB, 1TB
- Best BuyRead on 28 Aug 2026 and not re-read since: Best Buy blocks our automated checks, so this number is old and we say so rather than imply it is current. It was $1,999 under ASUS's own store once ASUS raised its price in September.in stockchecked 33 days ago$3,999.99
- ASUSlist pricein stockchecked 2 hours ago$5,999
- Mac mini, M6, 16GB$1,099Won't fit−73GB headroom
- Mac mini, M6, 24GB$1,299Won't fit−65GB headroom
- Mac mini, M6, 32GB$1,499Won't fit−57GB headroom
- Mac mini, M5 Pro, 24GB$1,699Won't fit−65GB headroom
- Mac mini, M5 Pro, 48GB$2,499Won't fit−42GB headroom
- Mac Studio, M5 Max, 36GB$2,499Won't fit−53GB headroom
- Mac mini, M5 Pro, 64GB$2,899Won't fit−28GB headroom
- GeForce RTX 5090, 32GB$2,999.99+hostWon't fit−56GB headroom
- Won't fit−28GB headroom
- Mac Studio, M5 Max, 48GB$3,099Won't fit−42GB headroom
- Mac Studio, M5 Max, 64GB$3,499Won't fit−28GB headroom
- Fits34GB headroomest. 42.8 tok/s
- Tight2GB headroomest. 128 tok/s
- Fits34GB headroomest. 85.6 tok/s
- Fits34GB headroomest. 42.8 tok/s
- Fits34GB headroomest. 96.3 tok/s
- Tight2GB headroomest. 188 tok/s
- Fits162GB headroomest. 128 tok/s
- Fits162GB headroomest. 188 tok/s
- Fits418GB headroomest. 188 tok/s
| Config | Memory | Price | $/GB | Fit | Headroom | Est. speed |
|---|---|---|---|---|---|---|
| Mac mini, M6 | 16GB | $1,099 | $69 | Won't fit | −73GB | — |
| Mac mini, M6 | 24GB | $1,299 | $54 | Won't fit | −65GB | — |
| Mac mini, M6 | 32GB | $1,499 | $47 | Won't fit | −57GB | — |
| Mac mini, M5 Pro | 24GB | $1,699 | $71 | Won't fit | −65GB | — |
| Mac mini, M5 Pro | 48GB | $2,499 | $52 | Won't fit | −42GB | — |
| Mac Studio, M5 Max | 36GB | $2,499 | $69 | Won't fit | −53GB | — |
| Mac mini, M5 Pro | 64GB | $2,899 | $45 | Won't fit | −28GB | — |
| GeForce RTX 5090 | 32GB | $2,999.99+host | $94 | Won't fit | −56GB | — |
| Mac Studio, M4 Max, refurb | 64GB | $3,049 | $48 | Won't fit | −28GB | — |
| Mac Studio, M5 Max | 48GB | $3,099 | $65 | Won't fit | −42GB | — |
| Mac Studio, M5 Max | 64GB | $3,499 | $55 | Won't fit | −28GB | — |
| Ascent GX10, GB10 Grace Blackwell | 128GB | $3,999.99 | $31 | Fits | 34GB | 42.8 tok/s |
| Mac Studio, M3 Ultra, refurb | 96GB | $4,329 | $45 | Tight | 2GB | 128 tok/s |
| Mac Studio, M4 Max, refurb | 128GB | $4,409 | $34 | Fits | 34GB | 85.6 tok/s |
| DGX Spark, GB10 Grace Blackwell | 128GB | $4,699 | $37 | Fits | 34GB | 42.8 tok/s |
| Mac Studio, M5 Max | 128GB | $5,099 | $40 | Fits | 34GB | 96.3 tok/s |
| Mac Studio, M5 Ultra | 96GB | $5,499 | $57 | Tight | 2GB | 188 tok/s |
| Mac Studio, M3 Ultra, refurb | 256GB | $7,729 | $30 | Fits | 162GB | 128 tok/s |
| Mac Studio, M5 Ultra | 256GB | $9,499 | $37 | Fits | 162GB | 188 tok/s |
| Mac Studio, M5 Ultra | 512GB | — | — | Fits | 418GB | 188 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 = 80B × 1.063 B/param (8.5 bits) + 0.025 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.5 ÷ (3B active × 1.063 B/param) = 3.2 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 Ascent GX10 GB10 Grace Blackwell 128GB runs
- Qwen3-235B-A22BQ3_K_M, est. 12.7 tok/s
- MiniMax-M2Q3_K_M, est. 27.9 tok/s
- Mistral-Large-2411Q6_K, est. 1.8 tok/s
- gpt-oss-120bQ6_K, est. 32.6 tok/s
- Llama-4-Scout-17B-16EQ8_0, est. 7.6 tok/s
- Qwen3-Coder-NextQ8_0, est. 42.8 tok/s
Best quant that still fits at 8K context, largest model first. Full list on the Config page.
Bandwidth 273GB/s, usable memory 120GB of 128GB.