What Can the AMD RX 7900 XT Run?

20 GB · compatibility with 168 AI models

Home › GPUs › AMD RX 7900 XT

Short answer: the AMD RX 7900 XT (20 GB) can run 165 of 168 AI image, video, 3D and audio models we track. Best picks: MiniMax H3 FL2VA, MiniMax H3 Ref2VA, Qwen-Image 2.1, Z-Image Turbo, WAN 2.2 T2V A14B (MoE).

S PerfectA GreatB GoodC UsableD BarelyF Can't Run
13 models77 models60 models12 models3 models3 models
Open the AMD RX 7900 XT in the checker

AMD RX 7900 XT specs for AI

BrandAMD
SeriesRX 7000
VRAM20 GB
Architecturerdna3
FP16 compute25.6 TFLOPS
Memory bandwidth800 GB/s
FP8 supportNo
Released2022
Launch price$799

AI models on the AMD RX 7900 XT

Each model uses the best format that fits in 20 GB. How ratings work.

Text to Image

ModelRatingRuns asVRAM usedSpeed
Stable Diffusion 1.5A GreatFP164 GBFast
Stable Diffusion 1.4A GreatFP164 GBFast
Stable Diffusion 2.1A GreatFP165 GBFast
Stable Diffusion 2.0A GreatFP165 GBFast
PixArt-Alpha 600MA GreatFP166 GBFast
SDXL Base 1.0A GreatFP167 GBFast
SDXL TurboA GreatFP167 GBFast
Stable Diffusion 3.5 Medium (2.5B)A GreatFP168 GBFast
Animagine XL 3.1A GreatFP167 GBFast
Pony Diffusion v6 XLA GreatFP167 GBFast
NoobAI XLA GreatFP167 GBFast
Juggernaut XL RagnarokA GreatFP167 GBFast
SDXL Lightning 4-stepA GreatFP167 GBFast
DreamShaper XLA GreatFP167 GBFast
Illustrious XL v1.0A GreatFP167 GBFast
Playground v2.5 2.6BA GreatFP167 GBFast
RealVisXL v5.0A GreatFP167 GBFast
Epicrealism XLA GreatFP167 GBFast
PixArt-Sigma 900MA GreatFP167 GBFast
Stable Diffusion 3 Medium (2B)A GreatFP167 GBFast
SDXL Hyper 1-stepA GreatFP167 GBFast
HunyuanDiT 1.5BA GreatFP168 GBFast
Z-Image TurboA GreatFP1610 GBFast
MOSS-Image 5BB GoodFP1612 GBFast
Flux.1 LiteB GoodFP1612 GBFast
AuraFlow 0.3B GoodFP1612 GBFast
Flux.2 Klein 4BB GoodFP1613 GBFast
Qwen-Image 2.1B GoodFP1614 GBFast
Krea 2 TurboB GoodGGUF Q48 GBModerate
Flux.1 Dev 12BB GoodGGUF Q48 GBModerate
Flux.1 Schnell 12BB GoodGGUF Q48 GBModerate
Z-Image BaseB GoodFP1614 GBFast
Ming-Image 0.1 DesignB GoodFP1614 GBFast
FLUX.1 Krea DevB GoodGGUF Q47 GBModerate
KolorsB GoodFP1614 GBFast
ERNIE-Image 8BB GoodFP1616 GBFast
HiDream-O1-Image 8.8BB GoodFP1616 GBFast
Qwen-Image 2512 (20B)B GoodGGUF Q412 GBModerate
FLUX.2 Klein 9BB GoodFP1618 GBModerate
Stable Diffusion 3.5 Large (8B)B GoodFP1618 GBModerate
Stable Diffusion 3.5 Large TurboB GoodFP1618 GBModerate
Chroma1 HDB GoodFP1618 GBModerate
Ideogram 4.0B GoodFP1619 GBModerate
Flux.2 Dev 32BC UsableGGUF Q418 GBSlow
HunyuanImage 3.0C UsableFP1620 GBModerate
Seedream 3.0C UsableFP1616 GBFast
GLM-ImageC UsableFP1616 GBFast

Image to Image

ModelRatingRuns asVRAM usedSpeed
SD 1.5 img2imgA GreatFP164 GBFast
InstructPix2PixA GreatFP165 GBFast
SDXL img2imgA GreatFP167 GBFast
IP-Adapter SDXLA GreatFP169 GBFast
IP-Adapter FaceID SDXLA GreatFP1610 GBFast
Mage-Flow 4BA GreatFP1610 GBFast
Boogu-Image EditB GoodGGUF Q46 GBModerate
Flux.1 Dev img2img 12BB GoodGGUF Q48 GBModerate
Flux Kontext DevB GoodGGUF Q48 GBModerate
LongCat-Image EditB GoodFP1615 GBFast
Qwen-Image-Edit 2511B GoodGGUF Q412 GBModerate

Inpainting

ModelRatingRuns asVRAM usedSpeed
SD 1.5 InpaintingA GreatFP164 GBFast
SDXL InpaintingA GreatFP167 GBFast
BrushNetA GreatFP168 GBFast
PowerPaint v2.1A GreatFP167 GBFast
Flux.1 Fill DevB GoodGGUF Q48 GBModerate

ControlNet

ModelRatingRuns asVRAM usedSpeed
ControlNet SD 1.5A GreatFP165 GBFast
T2I-Adapter SDXLA GreatFP168 GBFast
ControlNet SDXLA GreatFP169 GBFast
ControlNet FluxB GoodGGUF Q410 GBModerate

Video

ModelRatingRuns asVRAM usedSpeed
WAN 2.1 T2V 1.3BA GreatFP166 GBFast
AnimateDiff LightningA GreatFP166 GBFast
AnimateDiff v3 (SD 1.5)A GreatFP166 GBFast
ZeroScope v2 576wA GreatFP166 GBFast
AnimateDiff V2VA GreatFP166 GBFast
AnimateDiff I2VA GreatFP166 GBFast
Stable Video Diffusion 1.1A GreatFP168 GBFast
CogVideoX 2BA GreatFP168 GBFast
ModelScope T2VA GreatFP168 GBFast
TokenFlowA GreatFP168 GBFast
LatteA GreatFP168 GBFast
SEINEA GreatFP168 GBFast
Kandinsky 6.0 Video Lite 3BA GreatFP169 GBFast
LTX-Video 0.9 (2B)A GreatFP1610 GBFast
LTX-Video I2VA GreatFP1610 GBFast
Stable Video Diffusion XT 1.1 (25 frames)A GreatFP1610 GBFast
DynamiCrafter 512A GreatFP1610 GBFast
WAN 2.2 TI2V 5BB GoodFP1612 GBFast
AnimateDiff SDXLB GoodFP1612 GBFast
WAN 2.2 T2V A14B (MoE)B GoodGGUF Q48 GBModerate
WAN 2.2 I2V A14BB GoodGGUF Q48 GBModerate
WAN 2.1 T2V 14BB GoodGGUF Q48 GBModerate
WAN 2.1 I2V 480P 14BB GoodGGUF Q48 GBModerate
WAN 2.1 V2VB GoodGGUF Q48 GBModerate
CogVideoX 5BB GoodFP1614 GBFast
CogVideoX 5B I2VB GoodFP1614 GBFast
SkyReels V1B GoodFP1614 GBFast
SkyReels V1 I2VB GoodFP1614 GBFast
SkyReels V3 R2V 14BB GoodGGUF Q49 GBModerate
Wan 2.2 S2V 14BB GoodGGUF Q410 GBModerate
WAN 2.1 I2V 720P 14BB GoodGGUF Q410 GBModerate
Wan-Animate-2 14BB GoodGGUF Q410 GBModerate
SCAIL-2 14BB GoodGGUF Q410 GBModerate
Wan-Dancer 14BB GoodGGUF Q410 GBModerate
HuMo 17BB GoodGGUF Q410 GBModerate
Wan 2.2 Animate 14BB GoodGGUF Q411 GBModerate
LTX-2 19BB GoodGGUF Q411 GBModerate
MiniMax H3 FL2VAB GoodGGUF Q412 GBModerate
MiniMax H3 Ref2VAB GoodGGUF Q412 GBModerate
LTX-2.5 22BB GoodGGUF Q413 GBModerate
LTX-2.3 22BB GoodGGUF Q413 GBModerate
DynamiCrafter 1024B GoodFP1618 GBModerate
Ovi 1.1B GoodOffload5 GBVery Slow
LongCat-Video 13.6BB GoodOffload6 GBVery Slow
Open-Sora 2.0 (11B)C UsableOffload12 GBVery Slow
Waver 1.0C UsableFP1612 GBFast
HunyuanVideo T2V v1.5C UsableGGUF Q48 GBModerate
HunyuanVideo I2V v1.5C UsableGGUF Q48 GBModerate
HunyuanVideo 13B T2VC UsableGGUF Q48 GBModerate
HunyuanVideo I2VC UsableGGUF Q48 GBModerate
FramePack I2V HY FP8C UsableOffload8 GBVery Slow
LTX-Video 13BD BarelyOffload12 GBVery Slow
MAGI-1D BarelyOffload12 GBVery Slow
Mochi 1 10BD BarelyOffload16 GBVery Slow
MOVAF Can't Run—needs 29 GB—

Upscaler

ModelRatingRuns asVRAM usedSpeed
4x-UltraSharpS PerfectFP320.3 GBFast
4x-RemacriS PerfectFP320.3 GBFast
SwinIRS PerfectFP320.5 GBFast
RealESRGAN x4plusS PerfectFP320.3 GBFast
HATS PerfectFP320.5 GBFast
DATS PerfectFP320.5 GBFast
ScuNETS PerfectFP320.3 GBFast
LDSRA GreatFP164 GBFast
SeedVR2 3BA GreatFP169 GBFast
SeedVR2 7BB GoodFP1616 GBFast

Face/Restore

ModelRatingRuns asVRAM usedSpeed
GFPGAN v1.4S PerfectFP320.3 GBFast
CodeFormer v0.1S PerfectFP320.3 GBFast

VAE

ModelRatingRuns asVRAM usedSpeed
SD 1.5 VAE ft-mseS PerfectFP320.2 GBFast
Flux VAES PerfectFP320.3 GBFast
SDXL VAES PerfectFP320.3 GBFast
TAESD / TAESDXLS PerfectFP320.1 GBFast

3D

ModelRatingRuns asVRAM usedSpeed
TripoSRA GreatFP164 GBFast
Shap-EA GreatFP164 GBFast
InstantMeshA GreatFP166 GBFast
Zero123++ v1.2A GreatFP166 GBFast
LGMA GreatFP168 GBFast
One-2-3-45A GreatFP168 GBFast
Hunyuan3D 2.1A GreatFP1610 GBFast
SV3DA GreatFP1610 GBFast
Pixal3DB GoodFP1614 GBFast
Hunyuan3D 2.0C UsableFP1612 GBFast
TRELLIS.2 4BF Can't Run—needs 12 GB—
SAM 3D ObjectsF Can't Run—needs 32 GB—

Audio

ModelRatingRuns asVRAM usedSpeed
MusicGen Small (300M)A GreatFP162 GBFast
Whisper Large v3A GreatFP164 GBFast
Chatterbox MultilingualA GreatFP164 GBFast
XTTS v2A GreatFP163 GBFast
Stable Audio 3 Small-SFXA GreatFP163 GBFast
MMAudio Large 44kA GreatFP164 GBFast
Stable Audio OpenA GreatFP164 GBFast
Qwen3-TTS 1.7BA GreatFP165 GBFast
MusicGen Medium (1.5B)A GreatFP166 GBFast
IndexTTS 2.5A GreatFP166 GBFast
ACE-Step 1.5 TurboA GreatFP168 GBFast
VibeVoice 1.5BA GreatFP167 GBFast
BarkA GreatFP168 GBFast
Stable Audio 3 MediumA GreatFP167 GBFast
MusicGen Large (3.3B)B GoodFP1612 GBFast
Fish Audio S2 Pro 4BB GoodGGUF Q47 GBModerate
YuE2 3BB GoodOffload8 GBVery Slow
MiniMax Music 3B GoodOffload8 GBVery Slow

Frequently asked questions

How much VRAM does the AMD RX 7900 XT have?

The AMD RX 7900 XT has 20 GB of VRAM with 800 GB/s of memory bandwidth.

What AI models can the AMD RX 7900 XT run?

It can run 165 of the 168 AI models we track: 13 perfectly (S), 77 great (A), 60 good (B), 12 usable (C) and 3 barely (D). 3 need more VRAM than it has.

Can the AMD RX 7900 XT run MiniMax H3 FL2VA?

Yes, in GGUF Q4 (rating B, good).

Can the AMD RX 7900 XT run MiniMax H3 Ref2VA?

Yes, in GGUF Q4 (rating B, good).

Can the AMD RX 7900 XT run Qwen-Image 2.1?

Yes, in FP16 (rating B, good).

Can the AMD RX 7900 XT run Z-Image Turbo?

Yes, in FP16 (rating A, great).

Similar AMD GPUs