GPUs & Macs › AMD Radeon RX 7900 GRE
AMD Radeon RX 7900 GRE: local LLM speed
Llama 2 7B (Q4_0) generates 116 tokens/s on the AMD Radeon RX 7900 GRE (Vulkan), and processes prompts at 2,336 tokens/s.
Measured results
| Model | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Llama 2 7B | 116 | 2,336 | Q4_0 | Vulkan FA off | tg128 | llama.cpp #10879 details |
| Llama 2 7B | 110 | 2,251 | Q4_0 | Vulkan FA on | tg128 | llama.cpp #10879 details |
| Llama 2 7B | 97.5 | 1,599 | Q4_0 | ROCm FA on | tg128 | llama.cpp #15021 details |
| Llama 2 7B | 96.1 | 1,457 | Q4_0 | ROCm FA off | tg128 | llama.cpp #15021 details |
| Llama 3.2 1B Instruct | 125 | 3,610 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
What fits on the AMD Radeon RX 7900 GRE
Estimates for this chip are scaled from a measured run: Llama 2 7B Q4_0 at 116 tok/s (Vulkan, source). Fit assumes the whole model in GPU memory with ~8K context.
| Model | Best quant that fits | File | Expected tok/s |
|---|---|---|---|
| Qwen3.8-27B | UD-IQ3_XXS | 10.9 GB | 40.6 estimate |
| gemma-4-12B-it | Q8_0 | 12.7 GB | 35.0 estimate |
| Qwen3.5-9B | Q8_0 | 9.5 GB | 46.6 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 99.0 estimate |
| Qwen3.6-35B-A3B | UD-IQ3_XXS | 13.2 GB | 196 estimate |
| gemma-4-26B-A4B-it | UD-IQ4_XS | 13.6 GB | 106 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q8_0 | 9.5 GB | 46.5 estimate |
| gemma-4-E4B-it | Q8_0 | 8.2 GB | 54.2 estimate |
| Muse-Glimmer-30B | UD-IQ3_XXS | 13.1 GB | 33.8 estimate |
| Qwen3-4B | Q8_0 | 4.3 GB | 104 estimate |
| gemma-4-31B-it | UD-IQ3_XXS | 11.8 GB | 37.5 estimate |
| gemma-4-12B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 6.7 GB | 66.1 estimate |
| gemma-3-1b-it | Q8_0 | 1.1 GB | 370 estimate |
| gemma-4-E2B-it | Q8_0 | 5.0 GB | 89.3 estimate |
Similar hardware
AMD Radeon RX 9070 XT · AMD Radeon RX 9070 · AMD Radeon RX 7800 XT · AMD Radeon RX 6900 XT · AMD Radeon RX 6800 · AMD Radeon VII · AMD Radeon RX 6800 XT · AMD Radeon RX 9060 XT