Tired of blurry fingers, noise artifacts, and messy, tangled node graphs for MiniMax H3? This workflow is built for high-performance, rock-solid stability, and long-form video generation.Designed specifically to run efficiently on 16 GB VRAM GPUs (like the RTX 5060 Ti) while delivering professional-grade results. UPDATE v1.2 Changes<span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">- Fixed Preview functionality. UPDATE v1.1 ChangesVisuals- Greatly simplified and cleaner layout for better readabilityFixes- Fixed first pass metadata/info handling- Fixed Single Pass mode (already patched in v1.0b)- Fixed Supersample / Upscale toggle logic- Resolved various minor bugs <h2 id="key-features">Key Features</h2><ul><li>True Dual-Pass Efficiency: Starts with an ultra-fast 0.21MP structural pass for motion and physics, followed by a 1MP latent refinement pass for crystal-clear fine details.</li><li>Pixel-Perfect Scaling: Precise mathematical aspect-ratio alignment between passes to avoid any stretching or blur.</li><li>Clean Patch Order: Solves hand and face jitter by maintaining mathematical headroom and utilizing a clean attention patch sequence (<code>ModelAttentionBackend</code> -&gt; <code>ModelSampling</code> -&gt; <code>Chunk FF</code> -&gt; <code>EasyCache</code>).</li><li>Motion Locked: Integrated with <code>BetterMoveV2</code> LoRA for seamless camera rotations and fast character movement.</li><li>Clean &amp; Modular: Fully annotated, neatly organized, and free of useless node clutter.</li></ul><h2 id="user-friendly-control-panel">User-Friendly Control Panel</h2>No need to dig into complex wire webs. Control your entire generation pipeline from a dedicated, clean control panel:<ul><li>Dual-Pass &amp; Upscale Toggles: Switch seamlessly between Latent Upscale and RTX/Supersampling upscalers depending on your GPU resources.</li><li>First Pass Ratio Control: Easily fine-tune your structural pass ratio (0.21 by default for optimal VRAM/speed balance).</li><li>NVIDIA RTX Post-Processing Suite: Built-in toggles for RTX Upscale, RTX Denoise, and RTX Deblur for instant video post-refinement.</li><li>Model &amp; LoRA Switches: Instant toggles for H3 Hybrid Model (FLV/REF), Turbo LoRAs, and NSFW LoRA.</li><li>Seamless Scene Extension (Continue Feature): Extend your video duration continuously with a single click. Currently, the extended video clip needs to be stitched manually in your video editor (Premiere, DaVinci, CapCut, etc.). Be sure to account for the overlapping frames for a perfect transition! (Full automated video concatenation will be integrated in a future update).</li><li>Audio &amp; Preview Controls: Toggle real-time generation previews or mute embedded audio.</li></ul>Performance Benchmark: Full 13.7s video generated in ~869 seconds (Latent Upscale Pass 1MP) on a single RTX 5060 Ti (16 GB). Required Models &amp; Custom Nodes<h3 id="required-models-custom-nodes-and-checkpoints">Required Models, Custom Nodes &amp; Checkpoints</h3><ul><li>VAE:<ul><li>Video VAE: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/vae/minimax_h3_video_vae_fp16.safetensors?download=true">minimax_h3_video_vae_fp16.safetensors</a></li><li>Audio VAE: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/vae/minimax_h3_audio_vae_fp32.safetensors?download=true">minimax_h3_audio_vae_fp32.safetensors</a></li></ul></li><li>Diffusion Models:<ul><li>Base FL: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/diffusion_models/minimax_h3_fl2va_pruned_fp8_scaled.safetensors?download=true">minimax_h3_fl2va_pruned_fp8_scaled.safetensors</a></li><li>Base REF: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/diffusion_models/minimax_h3_ref2va_pruned_fp8_scaled.safetensors?download=true">minimax_h3_ref2va_pruned_fp8_scaled.safetensors</a></li><li>Hybrid Alternatives: <a target="_blank" rel="ugc" href="https://civarchive.com/models/2851079/h3-eros-max">H3 Eros Max</a> | <a target="_blank" rel="ugc" href="https://civarchive.com/models/2877206/dasiwa-minimax-h3">DaSiWa MiniMax H3</a></li></ul></li><li>Text Encoders:<ul><li>Standard: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/text_encoders/qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors?download=true">qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors</a></li><li>Uncensored: <a target="_blank" rel="ugc" href="https://huggingface.co/sakamakismile/Qwen3-VL-32B-Heretic-MiniMax-H3-NVFP4/blob/main/qwen3vl_32b_heretic_minimax_h3_nvfp4.safetensors?download=true">qwen3vl_32b_heretic_minimax_h3_nvfp4.safetensors (NSFW)</a></li></ul></li><li>LoRAs:<ul><li>Turbo 4-Step FL: <a target="_blank" rel="ugc" href="https://huggingface.co/lightx2v/Minimax-h3-Turbo/resolve/main/minimax_h3_fl2v_turbo_4step_v1.2_768p_comfyui_bf16.safetensors?download=true">minimax_h3_fl2v_turbo_4step_v1.2_768p_comfyui_bf16.safetensors</a></li><li>Turbo 4-Step REF: <a target="_blank" rel="ugc" href="https://huggingface.co/lightx2v/Minimax-h3-Turbo/resolve/main/minimax_h3_ref2v_turbo_4step_v0.1_comfyui_bf16.safetensors?download=true">minimax_h3_ref2v_turbo_4step_v0.1_comfyui_bf16.safetensors</a></li><li>Ref Experimental: <a target="_blank" rel="ugc" href="https://huggingface.co/Kijai/MiniMax-H3-experimental/resolve/main/loras/minimax_h3_ref_lora_rank_256_bf16.safetensors?download=true">minimax_h3_ref_lora_rank_256_bf16</a></li><li>Motion &amp; Reasoning: <a target="_blank" rel="ugc" href="https://civarchive.com/models/2497207/h3-ltx-23-i2v-t2v-video-reasoning-lora-vbvr">H3 - LTX 2.3 VBVR</a> | <a target="_blank" rel="ugc" href="https://civarchive.com/models/2734359/better-motion-ltxminimax-h3">Better Motion LTX/Minimax H3</a></li><li>Style: <a target="_blank" rel="ugc" href="https://civarchive.com/models/2856467/mmh3-mystic-xxx">[MMH3] Mystic XXX (NSFW)</a></li></ul></li><li>Latent Upscaler Model:<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/LBH-123-AI/Minimax_h3_latent_Upscaler">Minimax_h3_latent_Upscaler</a></li></ul></li></ul><h3 id="custom-nodes">Custom Nodes</h3>Make sure these node suites are installed via ComfyUI Manager:<ul><li><a target="_blank" rel="ugc" href="https://github.com/CGlide/ComfyUI-CGlide">ComfyUI-CGlide (H3 Studio)</a></li><li><a target="_blank" rel="ugc" href="https://github.com/LBH-123-AI/Comfyui_Minimax_h3_latent_Upscaler">Comfyui_Minimax_h3_latent_Upscaler</a></li><li><a target="_blank" rel="ugc" href="https://github.com/whmc76/ComfyUI-NVIDIA-RTX-VSR-Pro">ComfyUI-NVIDIA-RTX-VSR-Pro</a></li><li><a target="_blank" rel="ugc" href="https://github.com/PixWizardry/ComfyUI-PIX_Nvidia_RTX_Deblur_Denoise">ComfyUI-PIX_Nvidia_RTX_Deblur_Denoise</a></li><li><a target="_blank" rel="ugc" href="https://github.com/Saganaki22/ComfyUI-sol-attn">ComfyUI-sol-attn</a></li><li>ComfyUI-KJNodes</li><li>ComfyUI-VideoHelperSuite</li><li>ComfyUI-LogicUtils</li></ul>

MICW - MiniMax H3 - Modular, Intuitive & Clean Workflow (Ref2va, Fl2va & Video Continuation)

UPDATE v1.2 ChangesFixes- Fixed Preview functionality.

Files with SHA256: 38ef43684dd9e49f33d13c2ae52cb6179eec368f41762c61f3ce25ac15f8a41d