Datasheet for local AI49 models98 GPUsData read 2026-09-25
No adsNo tracking
Check my GPU →

Can the RX 7900 XTX 24 GB run HiDream-I1 (Full)?

Image model · 17B24 GB GDDR6 · 384-bit · 960 GB/s · RDNA 3Data 2026-09-25
Runs well
Yes — and comfortably.

Download Q8_0 (18.7 GB). With the model's working memory it needs about 21.5 GB, leaving 2.5 GB spare on 24 GB. Quality: practically identical to the original.

Best fileQ8_0
File size18.7 GB
VRAM needed~21.5 GB
System RAM48 GB+
Memory map · RX 7900 XTX 24 GB21.5 GB / 24 GB
012 GB24 GB
Weights Q8_0 · 18.7 GBWorking memory · 2.0 GBReserve · 0.8 GBFree · 2.5 GB
calculated from real file sizes plus working memory. Not yet measured on this setup. How this works.

01What to download

PartFileFolderSize
Modelhidream-i1-full-Q8_0.gguf
Q8_0 · city96/HiDream-I1-Full-gguf
models/unet18.7 GBDownload →
Text encodert5xxl_fp8_e4m3fn_scaled.safetensors
T5-XXL FP8 scaled · Comfy-Org/HiDream-I1_ComfyUI
models/text_encoders5.2 GBDownload →
Text encoderclip_l_hidream.safetensors
CLIP-L (HiDream) · Comfy-Org/HiDream-I1_ComfyUI
models/text_encoders0.2 GBDownload →
Text encoderclip_g_hidream.safetensors
CLIP-G (HiDream) · Comfy-Org/HiDream-I1_ComfyUI
models/text_encoders1.4 GBDownload →
Text encoderllama_3.1_8b_instruct_fp8_scaled.safetensors
Llama 3.1 8B Instruct FP8 · Comfy-Org/HiDream-I1_ComfyUI
models/text_encoders9.1 GBDownload →
VAEae.safetensors
FLUX.1 VAE (ae) · Comfy-Org/HiDream-I1_ComfyUI
models/vae0.3 GBDownload →
Total download · keep about the same free on disk34.9 GB

Alternatives: T5-XXL FP16 (9.8 GB). The 16-bit text encoder is a little more faithful; the smaller one is chosen here because it loads faster and needs less RAM. Sizes read from Hugging Face (2026-09-25). “Download” links start the file directly; the file name links are the same files the official ComfyUI workflows use.

System RAM: 48 GB or more. ComfyUI keeps the model file, the text encoder and the VAE in system RAM and moves them to the GPU as needed. With this set of files that is about 34.9 GB, plus roughly 6 GB for Windows, ComfyUI and a browser: 40.9 GB in total. With less RAM it still runs, but Windows starts swapping to disk and loading gets very slow. calculated

02Every HiDream-I1 Full file on 24 GB

FileSizeNeededOn this cardQualityDownload
16-bit
SAFETENSORS · Comfy-Org
34.2 GB37.0 GB13.0 GB too bigthe original weightsHugging Face →
Q8_0 ←
GGUF · city96
18.7 GB21.5 GBfits · 2.5 GB sparepractically identical to the originalHugging Face →
FP8
SAFETENSORS · Comfy-Org
17.1 GB19.9 GBfits · 4.1 GB sparepractically identical to the originalHugging Face →
Q6_K
GGUF · city96
14.7 GB17.5 GBfits · 6.5 GB sparevery close to the originalHugging Face →
Q5_K_M
GGUF · city96
13.0 GB15.8 GBfits · 8.2 GB spareclose; small differences in fine detailHugging Face →
Q4_K_M
GGUF · city96
11.5 GB14.3 GBfits · 9.7 GB sparegood; some loss in fine detail and textHugging Face →
Q3_K_M
GGUF · city96
8.8 GB11.6 GBfits · 12.4 GB sparenoticeable loss of detailHugging Face →
Q2_K
GGUF · city96
6.6 GB9.4 GBfits · 14.6 GB spareheavy loss; a last resortHugging Face →

Sizes from the Hugging Face file listing, read 2026-09-25. “Needed” = file + 2 GB working memory for this model + 0.8 GB kept free for the system.

03The text encoder

CLIP-L + CLIP-G + T5-XXL + Llama 3.1 8B: 15.9 GB as FP8 (all four together: CLIP-L 0.25 + CLIP-G 1.39 + T5-XXL FP8 5.16 + Llama 3.1 8B FP8 9.08 GB). ComfyUI encodes the prompt first and can push the encoder out of VRAM before sampling, so it does not have to fit together with the model. On 24 GB the FP8 encoder fits on its own, so prompt encoding stays fast.

04About this GPU

ComfyUI supports AMD GPUs on Windows officially since v0.7.0 (January 2026), through ROCm. Memory works the same as on NVIDIA, so the fit verdicts apply. Some custom nodes are NVIDIA-only. RDNA 3 has no FP8 compute, so FP8 files save memory but give no speed-up.

RX 7900 XTX 24 GB: 24 GB GDDR6 · 384-bit · 960 GB/s · RDNA 3. Everything that runs on the RX 7900 XTX 24 GB →

05What more VRAM would change

Nothing to gain for this model: the RX 7900 XTX 24 GB already runs a top-quality file entirely in VRAM. More memory would only help with bigger images, longer clips or several models at once.

What changes from the RX 7900 XTX 24 GB to the RTX 5090 32 GB →

06Measured and reported results

Nobody has sent measured numbers for this pair yet. If you run HiDream-I1 Full on a RX 7900 XTX 24 GB, send your time per image and peak VRAM and it will appear here, credited.

07Questions

How much VRAM does HiDream-I1 Full need?

Around 9.4 GB with the smallest file (Q2_K) and 19.9 GB with an 8-bit file (FP8), counting working memory and a small system reserve. The full 16-bit file needs about 37.0 GB.

Which HiDream-I1 Full file should I download for the RX 7900 XTX 24 GB?

Q8_0 (18.7 GB) from city96/HiDream-I1-Full-gguf. Load it with Unet Loader (GGUF) from the ComfyUI-GGUF node pack.

Is FP8 faster than GGUF on the RX 7900 XTX 24 GB?

No. This GPU has no FP8 compute, so ComfyUI converts FP8 weights back before the maths. FP8 only saves memory here; Q8_0 GGUF is the closer-to-original 8-bit pick.

How much do I need to download for HiDream-I1 Full on the RX 7900 XTX 24 GB?

About 34.9 GB for the model file, text encoder and VAE listed on this page, and the same again free on disk. With a 24 GB card, 48 GB of system RAM or more is recommended.