Run Qwen3.5 397B A17B locally

Qwen3.5 397B A17B is 397B parameters (17B active per token — a mixture-of-experts, so it still needs room for all 397B in memory). Here's the VRAM it needs at each quantization, the smallest GPU that fits, and how a self-hosted box compares with the cheapest verified API.

VRAM by quantization

QuantizationVRAM neededFits onNotes
int4 (Q4)238.2 GB4× H100 (320 GB) ($7.96/hr)smallest footprint, minor quality loss
int8 (Q8)476.4 GB8× H100 (640 GB) ($15.92/hr)near-lossless
fp16 (full)952.8 GBmulti-node (>640 GB)reference quality

VRAM ≈ params × bytes/param × 1.2 (overhead). See the fullmethod and break-even calculator.

Self-host vs API

Cheapest API for Qwen3.5 397B A17B is $2.73 / 1M tokens(blended) via Alibaba. Running it yourself in int4 fits a4× H100 (320 GB) at $7.96/hr — about $5,810.8/month at 24/7. Those two lines cross at roughly 2.1B tokens/month: below that the API wins on cost, above it the dedicated GPU does (assuming you keep it busy). Tune your own volume in the break-even calculator.

Run it at home

No cloud account needed: Qwen3.5 397B A17B in int4 (238.2 GB) fits a Mac Studio M5 Ultra (256 GB unified), a $9,499 one-off buy. Amortized over 3 years that's about $0.3615/hr — hardware only, electricity aside. Against the Alibaba API at $2.73/1M, buying it pays for itself after roughly 3.5B tokens total. Below that the API is cheaper; a machine that mostly sits idle rarely earns back its price.

Get the weights

Quantized builds on the model hubs — links search live, so they track new builds as they appear:

License: open weights under Apache 2.0 · terms. Compare all providers on the qwen3.5-397b-a17b API page.