Can LLaDA2.1 mini run on the RX 7900 XT?
Yes. At 8K context the best fit is Q8_0, using about 17.1 GB of the 19.0 GB usable on the RX 7900 XT. Estimated speed ceiling: ~525 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|
| 2-bit (Q2_K) | 6.5 GB Fits | 6.7 GB Fits | 7.6 GB Fits | 1489 |
| 3-bit (Q3_K_M) | 8.3 GB Fits | 8.4 GB Fits | 9.3 GB Fits | 1142 |
| 4-bit (Q4_K_M) | 10.1 GB Fits | 10.3 GB Fits | 11.2 GB Fits | 913 |
| 5-bit (Q5_K_M) | 11.6 GB Fits | 11.8 GB Fits | 12.7 GB Fits | 784 |
| 6-bit (Q6_K) | 13.3 GB Fits | 13.4 GB Fits | 14.4 GB Fits | 681 |
| 8-bit (Q8_0) | 16.9 GB Fits | 17.1 GB Fits | 18.0 GB Tight | 525 |
| 16-bit (F16/BF16) | 31.1 GB No | 31.3 GB No | 32.2 GB No | — |
• real GGUF file size. Speed is a bandwidth ceiling (800 GB/s), not a benchmark. Methodology.
More
LLaDA2.1 mini on every GPU and Mac · Every model that runs on the RX 7900 XT · Other models that fit 24 GB VRAM · Try other settings in the calculator
Data: Hugging Face Hub (inclusionAI/LLaDA2.1-mini), AMD specs. Data as of 2026-11-17.