§01·tested · /gpus/rtx-4090

Recipes tested on RTX 4090

20 community-tested setups — recipes and guides whose author ran them on this exact card.

nvidia40 series24GB VRAM

multimodalbeginner20GB+recipe
Gemma 4 E4B on RTX 4090: Multimodal Inference via BF16 (with optional Q8_0 / Q4_K_M GGUF)
3dintermediate12GB+recipe
Waypoint 1.5 on RTX 4090: Real-Time Interactive World Model at 720p
videointermediate24GB+recipe
LightX2V on RTX 4090: 4-Step Text-to-Video with Distilled Wan2.1-14B
videointermediate24GB+recipe
Wan 2.2 TI2V-5B on RTX 4090: 720p Text/Image-to-Video in ComfyUI
imagebeginner20GB+recipe
Flux.2 Klein 4B on RTX 4090: BF16 Full-Resident 4-Step Text-to-Image via Diffusers or ComfyUI
imageintermediate16GB+recipe
Juggernaut Z on RTX 4090: Cinematic Photoreal Fine-Tune of Z-Image Base at BF16 via Diffusers or ComfyUI
llmbeginner16GB+recipe
gpt-oss 20B on RTX 4090: MXFP4 chat via Ollama or vLLM
llmbeginner10GB+recipe
DeepSeek-R1-Distill-Qwen-14B on RTX 4090 via Ollama Q4_K_M GGUF
llmintermediate18GB+recipe
Gemma 4 26B A4B-it on RTX 4090: Local Multimodal Chat via Q4_K_M GGUF + llama.cpp
llmintermediate22GB+recipe
Qwen3-32B on RTX 4090: UD-Q4_K_XL GGUF via llama.cpp
llmbeginner10GB+recipe
Qwen3-14B on RTX 4090: Q4_K_M GGUF via Ollama or llama.cpp
llmbeginner6GB+recipe
Qwen3-8B on RTX 4090: Q4_K_M GGUF via Ollama or llama.cpp
imageintermediate21GB+recipe
Qwen-Image on RTX 4090: 20B Text-to-Image via ComfyUI FP8 (Native Path)
imageintermediate18GB+recipe
LongCat-Image (base T2I) on RTX 4090: Bilingual 6B Text-to-Image via diffusers
imageintermediate24GB+recipe
Chroma1-Base (V48) on RTX 4090: Uncensored 8.9B FLUX.1-Schnell De-Distillation via Diffusers BF16
imageintermediate20GB+recipe
HiDream-O1-Image on RTX 4090: 2048×2048 Text-to-Image with BF16 in ComfyUI
imagebeginner16GB+recipe
Z-Image Turbo on RTX 4090: 8-Step 1024x1024 Text-to-Image at BF16 in ~2.3s with Diffusers or ComfyUI
llmbeginner10GB+recipe
Llama 3.1 8B on RTX 4090: Local Chat via llama.cpp + Unsloth UD-Q4_K_XL GGUF
videointermediate14GB+recipe
HunyuanVideo-1.5 on RTX 4090: 480p Step-Distilled Text-to-Video in ~75 Seconds
videointermediate24GB+recipe
Wan 2.2 T2V-A14B on RTX 4090: 720p text-to-video in ComfyUI with FP8 scaled weights

Recipes tested on RTX 4090

Gemma 4 E4B on RTX 4090: Multimodal Inference via BF16 (with optional Q8_0 / Q4_K_M GGUF)

Waypoint 1.5 on RTX 4090: Real-Time Interactive World Model at 720p

LightX2V on RTX 4090: 4-Step Text-to-Video with Distilled Wan2.1-14B

Wan 2.2 TI2V-5B on RTX 4090: 720p Text/Image-to-Video in ComfyUI

Flux.2 Klein 4B on RTX 4090: BF16 Full-Resident 4-Step Text-to-Image via Diffusers or ComfyUI

Juggernaut Z on RTX 4090: Cinematic Photoreal Fine-Tune of Z-Image Base at BF16 via Diffusers or ComfyUI

gpt-oss 20B on RTX 4090: MXFP4 chat via Ollama or vLLM

DeepSeek-R1-Distill-Qwen-14B on RTX 4090 via Ollama Q4_K_M GGUF

Gemma 4 26B A4B-it on RTX 4090: Local Multimodal Chat via Q4_K_M GGUF + llama.cpp

Qwen3-32B on RTX 4090: UD-Q4_K_XL GGUF via llama.cpp

Qwen3-14B on RTX 4090: Q4_K_M GGUF via Ollama or llama.cpp

Qwen3-8B on RTX 4090: Q4_K_M GGUF via Ollama or llama.cpp

Qwen-Image on RTX 4090: 20B Text-to-Image via ComfyUI FP8 (Native Path)

LongCat-Image (base T2I) on RTX 4090: Bilingual 6B Text-to-Image via diffusers

Chroma1-Base (V48) on RTX 4090: Uncensored 8.9B FLUX.1-Schnell De-Distillation via Diffusers BF16

HiDream-O1-Image on RTX 4090: 2048×2048 Text-to-Image with BF16 in ComfyUI

Z-Image Turbo on RTX 4090: 8-Step 1024x1024 Text-to-Image at BF16 in ~2.3s with Diffusers or ComfyUI

Llama 3.1 8B on RTX 4090: Local Chat via llama.cpp + Unsloth UD-Q4_K_XL GGUF

HunyuanVideo-1.5 on RTX 4090: 480p Step-Distilled Text-to-Video in ~75 Seconds

Wan 2.2 T2V-A14B on RTX 4090: 720p text-to-video in ComfyUI with FP8 scaled weights