AMD Radeon RX 7900 GRE
RDNA 3 'Golden Rabbit Edition' — 16 GB at 576 GB/s, between the 6800 XT and 7900 XT. ROCm officially supported. The current AMD value choice in the $500-600 tier. Comfortable for any 7B/13B model + 32B Q4 with light offload. ~110-145 tok/s on 7B Q4.
Affiliate disclosure: as an Amazon Associate and partner of other retailers, we earn from qualifying purchases. The verdict on this page is our editorial opinion; affiliate links never influence what we recommend.
Sub-scores sum to 455 / 1000. Headline = 455 × 0.70 (Estimated-confidence discount) = 319. This is an algorithmic performance-tier score — distinct from, and often lower than, the editorial “Our verdict” below, which weighs value and real-world fit (especially for hardware we haven’t measured yet). How scoring works →
Extrapolated from 576 GB/s bandwidth — 57.6 tok/s estimated. No measured benchmarks yet.
Plain-English: Comfortable at 14B and below — snappy enough for a coding agent.
Verdicts extrapolated from catalog VRAM + bandwidth + ecosystem flags. Hover any chip for the rationale. Want measured numbers? Submit your own run with runlocalai-bench --submit.
This card is for the operator who wants a single-GPU rig that handles 7B and 13B models at full speed, with room to run 32B Q4 with partial offload, all without fighting ROCm compatibility. At 576 GB/s, the 7900 GRE delivers 110-145 tok/s on 7B Q4 and 60-80 tok/s on 13B Q4, making it one of the fastest options for those sizes. The 16 GB VRAM is the hard ceiling: 32B Q4 (18 GB) requires offloading layers to system RAM, dropping to ~15-25 tok/s depending on CPU/RAM speed. 70B models are out of reach. ROCm support is official, but operators should expect occasional driver quirks and slower PyTorch kernel tuning compared to CUDA. Pass on this card if 32B+ models at full inference speed are the primary workload, or if the budget allows stretching to a used 7900 XTX (24 GB) for ~$100 more. At $549 street price, the 7900 GRE is the current AMD value king for local AI, offering the best price-to-performance ratio for 7B-13B workloads.
›Why this rating
The 7900 GRE earns an 8.5 for delivering excellent throughput on 7B and 13B models at a competitive price, with official ROCm support. It loses points due to the 16 GB VRAM ceiling limiting larger model viability and the ongoing software maturity gap versus NVIDIA.
Overview
RDNA 3 'Golden Rabbit Edition' — 16 GB at 576 GB/s, between the 6800 XT and 7900 XT. ROCm officially supported. The current AMD value choice in the $500-600 tier. Comfortable for any 7B/13B model + 32B Q4 with light offload. ~110-145 tok/s on 7B Q4.
Some links above are affiliate links. We may earn a commission at no extra cost to you. How we make money.
Specs
| VRAM | 16 GB |
| Power draw (peak) | 260 W |
| Released | 2024 |
| MSRP | $549 |
| Backends | ROCm Vulkan |
Models that fit
Open-weight models small enough to run on AMD Radeon RX 7900 GRE with usable context.
Frequently asked
What models can AMD Radeon RX 7900 GRE run?
Does AMD Radeon RX 7900 GRE support CUDA?
How much does AMD Radeon RX 7900 GRE cost?
Where next?
Reviewed by RunLocalAI Editorial. See our editorial policy for how we research and verify hardware specifications.