從真實影片案例學習 MiniMax H3 提示詞

NookWright｜AIイラストと動画 · @nookwright

A creator animates a FLUX 3 Image Klein still into a 15-second H3 clip, including a cat sniffing a finger.

kz@AI術師【サブ趣味垢】 · @OpenPicGallery2

An app PR video delegates asset creation, H3 generation through ComfyUI and editing to Codex and Claude.

AI Bard Guild · @IsekaiBardGuild

H3 front and side motion videos guide Opus rigging for a Tripo3D character, then the rig runs in a Unity map.

紅 · @eikoung

A same-seed H3 Turbo test uses first, middle and last style guides and publishes API graphs with observed transition limits.

Robin Delta · @heyrobinai

A HyperFlow tutorial pairs H3 demo clips with local CLI instructions and a same-seed baseline comparison option.

ComfyUI Dev · @ComfyUIdev

A ComfyUI Dev video demonstrates creating a RefMod for MiniMax H3 with paired character and source footage.

TOSHIBoo ꧁としぶ꧂全資産クリプト · @TOSHIBoo_m

A drama-style commercial combines Antigravity and Argon with Colab A100, H3 video and Gemini TTS.

沉默的未知 · @hylwr

A local H3 user reports resolving reboots by reducing GPU power and Windows CPU performance limits.

すーさん · @su_nagomi

A website promo uses Claude Code with Renoise H3 video, Seed Audio narration and Lyria music, then adds captions and edits.

CALMDUST aka ちゃんこあ · @core_tan

A WebUI dev1.7 test compares SLA, Fast Spark, Spark and OFF on a four-step 1920×1088 H3 clip.

Enze Xie · @xieenze_jr

Sol-H3 Studio presents ComfyUI generation with native audio and reported RTX 5090 and DGX Spark timings.

keys 🧪 · @u1tra_instinct

A TensorFold 2.8.0 studio update combines H3 Ref2VA, motion previs, audio tools and local production controls.

ダルトワ★TV · @MireilleDartois

An H3 long-form test combines four segments, Sage Attention, Easy Cache and an int8 VAE with reported resolution timings.

AI Video Explorer · @rand_shao

A linked H3 and Wan 3.0 Prime comparison uses the same clay creature, reference still and twelve prompts.

Kamimoto(かみもと) · @sep_is_heim

An H3 creator reports a fourteen-second clip in 98.8 seconds on RTX 4070, with faster reruns but weak speech quality.

あきばあい | 秋葉原IT戦略研究所 · @akibaAIilab

A Vidubus workflow documents an FP16 video VAE workaround for reported Mac MPS failures with an int8 VAE.

akiraxx · @theeakiraxx

An H3 ComfyUI test compares twenty and eight steps, Spectrum acceleration and an AudioRefine pass for speech distortion.

デジタルクリエイティブAI · @DCAI_TW

A local Ref2VA comparison reports faster four-step Turbo generation alongside weaker reference control and visible noise.

フースイ@AIイラスト · @ai_t2i_231

A five-second H3 animation documents a conversational Codex workflow that assembled English instructions behind the scenes.

きゅうす のAIとつれづれ · @Ai_Quse

A creator records DGX Spark memory and power during a long H3 workload, separating measurement from speed speculation.

SixteenGB · @SixteenGB

A 16 GB RTX 5070 Ti eGPU runs an H3 Ref2VA talking-cat example with int8, sparse attention and a four-step LoRA.

Feicai · @zhu185178

A six-stage short-drama tutorial compares H3 clip pricing with the time and iteration cost of finishing an entire film.

码农暖爸 · @Delroy715

A same-seed H3 skin-rendering comparison combines two realism LoRAs, SigmaShift and a reported native 768p canvas.

kongo jun / 混合 順 · @jun_kongo

A creator publishes a ComfyUI H3 latent-outpainting tutorial with an illustrated example and a linked video.

Hao AI Lab · @haoailab

A two-DGX-Spark 768p demo accompanies QAD-v2, eight-step audio/video sampling and two NVFP4 decoder options.

Ueda · @megyo9

An H3 head-orbit video feeds MediaPipe triangulation, MakeHuman deformation, texture baking and VRM export.

John · @Marmot0519

A 111-second shared film combines Comfy H3 video with several rounds of Claude-assisted caption refinement.

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

An H3 R2V workflow update adds a Refer-LoRA toggle and advises disabling it for pure Ref2VA or hybrid models.

CALMDUST aka ちゃんこあ · @core_tan

A WebUI dev1.7 test reports 2560×1440, 243 frames and six steps with FlashGen FL2V LoRA, taking 432.8 seconds.

ダルトワ★TV · @MireilleDartois

A chained H3 test uses 22-frame latent handoffs and reports unwanted cuts and lights switching off between short segments.