Can granite 4.0 h tiny run on the RX 7600?
Yes. At 8K context the best fit is Q6_K, using about 6.0 GB of the 7.6 GB usable on the RX 7600. Estimated speed ceiling: ~239 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | 128K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|---|
| 2-bit (Q2_K) • | 3.0 GB Fits | 3.1 GB Fits | 3.2 GB Fits | 4.0 GB Fits | 527 |
| 3-bit (Q3_K_M) • | 3.7 GB Fits | 3.8 GB Fits | 4.0 GB Fits | 4.7 GB Fits | 407 |
| 4-bit (Q4_K_M) • | 4.6 GB Fits | 4.6 GB Fits | 4.8 GB Fits | 5.5 GB Fits | 322 |
| 5-bit (Q5_K_M) • | 5.2 GB Fits | 5.3 GB Fits | 5.5 GB Fits | 6.2 GB Fits | 276 |
| 6-bit (Q6_K) • | 5.9 GB Fits | 6.0 GB Fits | 6.2 GB Fits | 6.9 GB Tight | 239 |
| 8-bit (Q8_0) • | 7.5 GB Tight | 7.5 GB Tight | 7.7 GB No | 8.5 GB No | 185 |
| 16-bit (F16/BF16) • | 13.6 GB No | 13.6 GB No | 13.8 GB No | 14.5 GB No | — |
• real GGUF file size from ibm-granite/granite-4.0-h-tiny-GGUF. Speed is a bandwidth ceiling (288 GB/s), not a benchmark. Methodology.
More
granite 4.0 h tiny on every GPU and Mac · Every model that runs on the RX 7600 · Other models that fit 8 GB VRAM · Try other settings in the calculator
Data: Hugging Face Hub (ibm-granite/granite-4.0-h-tiny), AMD specs. Data as of 2026-11-17.