AMD Radeon RX 7900 XT, AI & Machine Learning Benchmarks & Specs

20GB · AI Score 6.2/100 · anchored estimate vs 51 measured cards

6.2 AI Score Includes estimates

We have not run AMD Radeon RX 7900 XT on our bench. These figures are anchored estimates, interpolated per workload against the 51 GPUs we did measure (confidence: low). On Llama 3.1 8B (Q4_K_M) AMD Radeon RX 7900 XT should deliver about 110 tokens/sec. Stepping up to Qwen3 32B it should hold roughly 27.1 tok/s. Llama 3.3 70B does not fit. It needs roughly 42GB and this card has 20GB. For image generation, SDXL should run near 2.8 it/s, while FLUX.1-dev won't fit at BF16 (needs ~26GB). 4 of the 12 workloads won't fit on 20GB at the tested precision, Llama 3.3 70B, FLUX.1-dev, FLUX.1 Kontext, Qwen-Image-Edit. We publish those as hard gates rather than quietly dropping to a smaller quant.

AI & Machine Learning benchmark results

Text Generation tok/s 5

Qwen3 4B174.6
Llama 3.1 8B110
Qwen2.5-Coder 14B62
Qwen3 32B27.1
WorkloadResultTelemetryData
Qwen3 4B174.6 tok/sestimatedEst.
Llama 3.1 8B110 tok/sestimatedEst.
Qwen2.5-Coder 14B62 tok/sestimatedEst.
Qwen3 32B27.1 tok/sestimatedEst.
Llama 3.3 70B✕ Won't fit VRAM-gated at this precisionEst.

Image Generation images/min 3

WorkloadResultTelemetryData
Stable Diffusion XL5.6 images/minestimatedEst.
Z-Image Turbo2.48 images/minestimatedEst.
FLUX.1 dev✕ Won't fit VRAM-gated at this precisionEst.

Image Editing images/min 2

WorkloadResultTelemetryData
FLUX.1 Kontext dev✕ Won't fit VRAM-gated at this precisionEst.
Qwen-Image-Edit✕ Won't fit VRAM-gated at this precisionEst.

Video Generation frames/s 2

WorkloadResultTelemetryData
LTX-Video (distilled)1.39 frames/sestimatedEst.
Wan 2.2 5B (720p)0.14 frames/sestimatedEst.
How this estimate is derived. This card hasn’t been through our bench yet, so its numbers are anchored estimates, interpolated from the 51 first-party measured cards (Bandwidth Theil-Sen ladder off NVIDIA anchors × vendor factor (LLM ×0.85, diffusion ×0.5); realistic VRAM floors). The VRAM “won’t fit” gates are exact, since they’re pure capacity limits. Confidence: low. Estimates are replaced with measured data as more silicon goes through the bench. Full methodology →

AMD Radeon RX 7900 XT specifications

ArchitectureRDNA 3 (Navi 31)
Stream processors5,376
VRAM20GB GDDR6
Memory bus320-bit
Memory bandwidth800 GB/s
Boost clock2,425 MHz
TDP315 W
Process5nm (chiplet)
InterfacePCIe 4.0 x16
Release date2022-12-13
Launch MSRP$899

Verdict, capable, but 20GB sets the ceiling

AMD Radeon RX 7900 XT scores 6.2/100, #41 of 102. It ran 8 of 12; 4 exceeded its 20GB. Figures are anchored estimates, not measurements, we flag that on every row.

Relative performance: where the AMD Radeon RX 7900 XT lands

100% = this card, AI & Machine Learning headline metric (AI Score). #10 of 61 desktop cards in this vertical.

GPURelative%AI Score
NVIDIA GeForce RTX 3090 Ti
137%8.5
NVIDIA Titan RTX
132%8.2
NVIDIA GeForce RTX 3090
124%7.7
AMD Radeon RX 7900 XTX
108%6.7
AMD Radeon RX 7900 XT
100%6.2
GeForce RTX 4080 Super
87%5.4
GeForce RTX 5080
79%4.9
NVIDIA GeForce RTX 4080
76%4.7
GeForce RTX 5070 Ti
76%4.7

Same card, other workloads: AMD Radeon RX 7900 XT Gaming benchmarks

← All AI & Machine Learning GPU rankings

The silicon

Transistors57,700 million
Process node5 nm
Fabricated byTSMC

Silicon figures from Wikipedia (CC BY-SA 4.0). Benchmarks on this page are our own. Compare every chip.

What this card can build

Whole-job timings, composed from our measured per-model results on this card.

WorkflowTimeEnergyBasis
24-frame storyboard10.6 minn/aestimate, 0 of 2 stages measured
Full codebase review16.1 minn/aestimate, 0 of 1 stage measured
60-second AI short film20.7 minn/aestimate, 0 of 3 stages measured
10 short social clips62.7 minn/aestimate, 0 of 3 stages measured

Can't run: 40-product photo shoot (needs FLUX.1 Kontext dev), 6-panel comic page (needs FLUX.1 dev), 20 long-form articles (needs Llama 3.3 70B), Character sheet, 12 poses (needs FLUX.1 dev), 100-photo restoration batch (needs FLUX.1 Kontext dev), 100-photo restore and enlarge (needs FLUX.1 Kontext dev).