Haocheng Xi 🚗 COLM · @HaochengXiUCB
FreeVideo DeltaNet Laptop Demo
A FreeVideo demonstration brings H3 with Video DeltaNet to local hardware and advertises an 8GB VRAM, 16GB RAM configuration.
Prompt not publishedView details
Verified prompts and original sources
MiniMax H3 Prompts With Real Video Examples
Study prompts published by real creators alongside their finished MiniMax H3 videos. See how scene direction, motion, camera work, continuity, and sound fit together.
Every example links back to its original source. Copy the published prompt as written or adapt its structure to your own scene.
706 verified examples
Browse the MiniMax H3 prompt collection
Browse MiniMax H3 examples by camera technique, storyboard, animation, reference control, music, dialogue or motion graphics. Published prompts and workflow-only examples are clearly identified.
214 examples found · Showing 31–60
MiniMax H3 ComfyUI & Local Workflow Examples
Browse supporting MiniMax H3 workflows for ComfyUI, local inference, acceleration, video continuation and post-production. These entries are workflow and comparison examples rather than a library of copyable prompts. Follow the original sources for author-published settings and tooling, and compare quality as well as reported runtime. Hardware tests depend on their specific setup. Keep the same input and generation conditions when evaluating a change, and check whether the source provides enough detail to reproduce it.
Kamimoto(かみもと) · @sep_is_heim
X2 VAE Portrait RTX 4070 Workflow
An RTX 4070 H3 workflow renders 544×960 and uses X2 VAE to reach 1088×1920, with a reported 83-second five-second clip.
Prompt not publishedView details
バスケットマン（基本無課金 多垢100連ガチャ） · @kaimakulink
Singularity and Fused RefDelta Comparison
A local comparison changes only the H3 variant, contrasting Singularity Ref2VA with a fused RefDelta Turbo configuration.
Prompt not publishedView details
ジャヴァノ🍀ボカロP@初音ミク「w/39」10/12アルバムリリース · @javawock7618
Native R2V Latent Upscale Workflow
An H3 R2V workflow update replaces VHS video nodes with native input/output and exposes latent-upscale step control.
Prompt not publishedView details
MiniMax H3えふのじ · @fnoji
Joined Scene Color Drift AE Repair
An After Effects comparison attempts to restore color and tonal range after connecting H3 scenes, showing corrected and untreated frames.
Prompt not publishedView details
MiniMax H3ジャヴァノ🍀ボカロP@初音ミク「w/39」10/12アルバムリリース · @javawock7618
Bridge Retake Between Video Cuts
An H3 Bridge workflow adds retakes alongside generating intermediate cuts and prepending a cut to an existing video.
Prompt not publishedView details
MiniMax H3JJ Eaton · @jayleaton
TensorFold RTX 5070 Ti Loader Benchmark
A Windows RTX 5070 Ti H3 benchmark compares a stock loader, TensorFold, sparse attention, Turbo LoRA and an int8 video VAE.
Prompt not publishedView details
A local H3 recipe animates Imagen 4 references at 1024×1024, then upscales and interpolates to 1536×1536 at 60 FPS.
MiniMax H3 (LightX2V FL2V Turbo)AIconia | アイコニア · @AiPhotorealGirl
Imagen 4 LightX2V RTX 4070 I2V Recipe
A local H3 recipe animates Imagen 4 references at 1024×1024, then upscales and interpolates to 1536×1536 at 60 FPS.
あいびぃ · @ivy432hz
Prerendered AITuber Stage
An AITuber prototype combines a coded 3D stage, prepared H3 clips, GPT Image 2 artwork and Irodori TTS.
Prompt not publishedView details
鈴木憂一 | Highdrama · @yu_ichi_suzuki
PDMD RTX 5090 Real-Time Tradeoffs
A two-step H3 experiment trades resolution, frame rate and stability for continuous generation on one RTX 5090.
🦠🦠☠️Rezo☠️🦠🦠 · @baharezo
Dizzy LongShot Study
An author shares a 112-second character study identified as a MiniMax H3 LongShot experiment.
Prompt not publishedView details
Kamimoto(かみもと) · @sep_is_heim
RTX 4070 INT8 X2 Parallel Decode
An updated H3 workflow combines partial INT8 X2 VAE, parallel decode/NVENC and RTX 4070 FFN tuning.
Prompt not publishedView details
えふのじ · @fnoji
LTX SDR to HDR H3 Postprocessing
An H3 clip is processed through LTX 2.5 SDR-to-HDR, with author-reported reduction in noise and flicker.
Prompt not publishedView details
角煮星丸 · @_3912657840
MiniMaxDesign ZIP Batch Speech
A creator packages images and Suno speech into a ZIP for agent-driven H3 clips, then adds captions and editing with Codex.
Prompt not publishedView details
🟢DeepForest_Studio_AI🟢 · @DeepForest_Inc
DMAD Four Eight Twelve Step Comparison
A four-panel H3 comparison contrasts DMAD at 4, 8 and 12 steps with a 25-step base render at 768×1376.
Prompt not publishedView details
MiniMax H3ためしばこ · @tameshibako_tc
Cloud Resolution and Time Tradeoff
A five-second H3 test on a 96GB cloud GPU compares about 42 seconds at 864×480 with 34 minutes at 2720×1536.
Prompt not publishedView details
keys 🧪 · @u1tra_instinct
Mac TensorFold Studio Image and Audio
A Mac Silicon studio combines Qwen Image 2.1 with an H3 MLX implementation and draft or high-quality options.
Prompt not publishedView details
ささゆう｜MiniMax H3・AI撮影スタジオKonte開発 · @forticodesign
KONTE Repeatable Cast and Prop Production
A twenty-commercial exercise reuses cast and props through a structured production flow directed with Claude.
Prompt not publishedView details
鈴木憂一 | Highdrama · @yu_ichi_suzuki
Seven Second PDMD Audio Pipeline
A 5090 and three Spark-compatible systems pipeline seven-second voiced clips, using Irodori audio as the H3 guide.
Prompt not publishedView details
AiDeNier（アイデニール） · @de_nier_ai
Eight GB Melting Output Test
An eight-second H3 output demonstrates an 8GB VRAM run while the author explicitly notes melting artifacts.
Prompt not publishedView details
新手老兵🇭🇰的AI世界 · @Langby2
Model Adapted Direction Comparison
A Seedance, WAN and H3 comparison uses separately adapted prompts and reports H3 foot and travel-direction errors.
Prompt not publishedView details
MiniMax H3遠野の世界 · @tonoseka
R2V Face Repair with AE Tracking
A low-quality local R2V test uses reference markers and After Effects tracking to attach a replacement face.
Prompt not publishedView details
Nobu-Kobayashi : Generative AI Technology · @nyaa_toraneko
Local 43-Second Continuity Test
A local H3 creator reports a connected 43-second movie, with an attached clip of about 43.4 seconds.
Prompt not publishedView details
ゆずき · @uzuki425
Dual GPU Realtime Lip-Sync Budget
A creator reports 7.3-second lip-sync clips generated in 7.0 seconds at 352×640 using a 5090 plus an 8GB-class GPU.
Prompt not publishedView details
Matt · @MrCollison
H3 Max Neural Motion WebGPU Playground
Nine H3 Max videos feed neurally matched character animations in a Three.js WebGPU playground with custom IK and blending.
Prompt not publishedView details
An RTX 4070 creator tests Veda against an existing four-step INT8, SLA and X2 VAE local H3 setup.
MiniMax H3Kamimoto(かみもと) · @sep_is_heim
Veda RTX 4070 Acceleration Test
An RTX 4070 creator tests Veda against an existing four-step INT8, SLA and X2 VAE local H3 setup.
Prompt not publishedView details
小此木 · @okonogiret
Turbo8 Motion Lab Quality-Time Tradeoff
Adding Motion Lab to a Turbo8 workflow slightly reduces breakup but increases reported runtime from about 11 to 20 minutes.
Prompt not publishedView details
Masa/マーケティング×生成AIの「株式会社ユーダイモニア」代表 · @psychonurseblog
Pet T-Rex RunPod Minute Production
A fictional pet T-Rex film combines RunPod H3 visuals, GPT-Image keyframes, external TTS and sound effects in a one-minute production.
Prompt not publishedView details
MiniMax H3あいもり · @AIgazouForest
RTX 3060 Maid Two-Stage Interpolation
A 3060 12GB/64GB RAM test reports about five minutes for a maid clip, followed by 24-to-48fps interpolation.
Prompt not publishedView details
MiniMax H3えーてぃーえふ · @ATFzwei
Maestro WebUI Local Multimodal Setup
A local Maestro user shows H3 WebUI setup with in-app acceleration LoRA installation and image/voice tools.
Prompt not publishedView details
Use the collection
How to Adapt MiniMax H3 Prompts
Treat each example as a reusable structure. Keep the parts that control motion and continuity, then replace the creative details that belong to the original scene.
Find the closest motion pattern
Browse a category or search by subject, camera technique, creator or model. Choose an example whose motion and structure match your intended video.
Copy the published prompt
Copy the published text when available, checking whether it is a full prompt or a fragment. For examples without a prompt, study the source and adaptation notes.
Replace scene-specific details
Swap in your own subject and setting, provide the required reference media, and review every constraint before generating.
Use a MiniMax H3 Prompt in the Video Generator
Copy a verified example, adapt the creative details, and generate your own video with MiniMax H3.