Can Inkling run on the RX 9070?
No. Even at 2-bit with 8K context, Inkling needs about 335 GB, more than the 15.2 GB usable on the RX 9070.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | 128K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|---|
| 2-bit (Q2_K) | 334 GB No | 335 GB No | 342 GB No | 366 GB No | — |
| 3-bit (Q3_K_M) | 435 GB No | 436 GB No | 443 GB No | 467 GB No | — |
| 4-bit (Q4_K_M) | 544 GB No | 545 GB No | 551 GB No | 576 GB No | — |
| 5-bit (Q5_K_M) | 634 GB No | 635 GB No | 641 GB No | 666 GB No | — |
| 6-bit (Q6_K) | 729 GB No | 730 GB No | 736 GB No | 761 GB No | — |
| 8-bit (Q8_0) • | 800 GB No | 801 GB No | 807 GB No | 832 GB No | — |
| 16-bit (F16/BF16) • | 1766 GB No | 1767 GB No | 1773 GB No | 1798 GB No | — |
• real GGUF file size from unsloth/inkling-GGUF. Speed is a bandwidth ceiling (640 GB/s), not a benchmark. Methodology.
Hardware that does fit it
More
Inkling on every GPU and Mac · Every model that runs on the RX 9070 · Other models that fit 16 GB VRAM · Try other settings in the calculator
Data: Hugging Face Hub (thinkingmachines/Inkling), AMD specs. Data as of 2026-11-17.