📈 推特热点 · AI 提示词趋势

数据区间 2026-07-02 ~ 2026-07-31 · 每日自动更新 · 更新于 2026-08-01 01:26 UTC
追踪 X/Twitter 上 AI 图像生成生态的热门关键词——按天统计各模型、风格与技法在公开推文中的提及热度，反映当前创作者社区的关注趋势。取近期提及量最高的关键词绘制曲线。下方展开每个关键词的关联热词（从匹配推文中实时提取的共现高频话题），以及具体的代表性推文。

关键词热度趋势曲线

纵轴 = 每日相关推文提及次数（采样） · 横轴 = 日期（UTC） · 悬停查看具体数值 · 点击图例可显隐曲线

📊 今日趋势总结

🔥 热度最高

Flux（311）、Midjourney（255）、电影感（136）

🚀 上升最快（环比昨日）

Logo（▲100%）、Flux（▲78%）、提示词工程（▲67%）

关键词区间总提及最新一日环比关联热词（从推文中提取）
Flux31141▲ 78%black hole phase太空 formula log wigner function inward collapse language models
Midjourney2555▼ 50%grok imagine extra credits buy extra aiart synthart synthart aiartcommunity weekly pool
电影感1360▼ 100%attached参考 depth field color scheme 浅景深 scene selector selected scene
动漫1310—attached参考 scene selector selected scene exactly small real person spectator splash
GPT Image 211620%exactly small grok imagine sans serif right panel 全身 study exactly
提示词工程11210▲ 67%medical education console log claude code machine learning immune system database encryption
Nano Banana Pro942▼ 33%gemini pro grok imagine gemini flash stranger things things others product gemini
照片写实690—attached参考 depth field skin纹理 scene selector selected scene 浅景深
ComfyUI686—open source voice cloning exactly need come mind open sourced social media
Logo402▲ 100%console log attached参考 medical education phase太空 scene selector selected scene

🔍 关键词关联话题详情

Flux提及 311 次 · 最新 41 次
black hole20phase太空18formula log15wigner function12inward collapse12language models12
"@FluxtraMANTRA FLUX Points Leaderboard TOP 5 OMie Multiplier ->Top Tier 1.50x FLUX Points Founder Multiplier ->Tier 4 1.88x FLUX Points"
@ — DeFi & Crypto (🕉, 🏘️)
"FLUX 3 is one model trained for image, video, audio and action-prediction for robotics. FLUX-mimic is already running on factory robots at Audi doing wiring, sorting and assembly w"
@ — Kol Tregaskes
"@FluxOnChain1 Sweet dreams Flux"
@ — Witch-King
Midjourney提及 255 次 · 最新 5 次
grok imagine14extra credits11buy extra9aiart synthart8synthart aiartcommunity8weekly pool8
"Two new image models in two days! Seedream 5.0 Pro Muse Image I tested both on the same cinematic prompt against Midjourney and GPT Image 2. Midjourney is still the aesthetic bench"
@ — V
"QT with Something Purple #aiartcommunity #aiart #AIArtwork #aiartist #midjourney #midjourneyart #midjourneyartwork #aiDigitalArt"
@ — Sam Twito
"Coffee in Toronto #flickr SANDRO #ChristinaSaintMarché #artificialintelligence #ai #machinelearning #aiart #digitalart #technology #art #aiartcommunity #midjourney #datascience #ge"
@ — Saint Marche
电影感提及 136 次 · 最新 0 次
attached参考45depth field22color scheme21浅景深19scene selector19selected scene19
"How I Created This AI Cinematic Video - This video was created using Happy Horse 1.1 @HappyHorseATH with 9 Image References to maintain character consistency and cinematic storytel"
@ — BMX
"Here's my version Same concept but last one with humor Authentic live-action cinematic footage shot on real 35mm Kodak film as one uninterrupted continuous tracking shot through a "
@ — ANKIT PATEL 🇮🇳 | AI
"HIRING: AI Filmmaker / AI Video Editor 📍 Remote We’re looking for a creative AI Filmmaker / AI Video Editor to produce high-quality AI-generated videos from concept to final delive"
@ — REMOTE JOB WITH TESSY
动漫提及 131 次 · 最新 0 次
attached参考45scene selector19selected scene19exactly small18real person16spectator splash16
"Prompt of the Day: MILITARY ANIME COSPLAY 🎖️⚓✈️💜💚 Today’s Prompt of the Day drops your character directly into the world of a military anime. Choose a specific anime in the MILITAR"
@ — E-Va 💜💚
"Mission Log #279: 🌈 Cosmic Prism of Infinite Resonance 🎧 Every melody carries a different color. Every color carries a different emotion. Aoba listens to the heartbeat of the cosmo"
@ — Astro Creator(Yosuke Tanahashi)🌌
"@Maegatron3030 @BluJay614 @KivoxsWife @ExoDust7755 @VesperB_VT Anime “Dark Smile” Over-the-Shoulder Reaction Prompt Create a polished anime-style image using the provided character"
@ — Zedra x Ai Zhunter
GPT Image 2提及 116 次 · 最新 2 次
exactly small18grok imagine16sans serif13right panel13全身12study exactly12
"Want to make stylized sketches using GPT-Image2? It works for Nanobannana as well but its preferable to use GPT-Image 2. Use the prompt below! 👇 "A highly detailed black-and-white "
@ — Kinovi.AI
"Two new image models in two days! Seedream 5.0 Pro Muse Image I tested both on the same cinematic prompt against Midjourney and GPT Image 2. Midjourney is still the aesthetic bench"
@ — V
"Nobody has run this test yet, so I did 🥊 The 4 frontier image models — Nano Banana Pro, GPT-Image-2, ByteDance's Seedream 5 Pro (out YESTERDAY) and @metaai brand new Muse: given th"
@ — Alex Volkov
提示词工程提及 112 次 · 最新 10 次
medical education31console log30claude code19machine learning18immune system18database encryption18
"Prompt Engineering vs Fine-Tuning: Which Should You Choose? One question I see quite often is: Should you invest in better prompts, or should you fine-tune your model? My answer is"
@ — Bad Girl
"prompt engineering harness engineering loop engineering ? recursion engineering pointer engineering stack engineering queue engineering tree engineering inheritance engineering abs"
@ — Hamel Husain
"The AI jobs people want are not won by saying prompt engineering. They are won by building systems that retrieve, validate, automate, and produce value."
@ — The Jobfather ® 🇯🇲🇨🇦🇬🇧
Nano Banana Pro提及 94 次 · 最新 2 次
gemini pro9grok imagine9gemini flash8stranger things7things others7product gemini7
"Nobody has run this test yet, so I did 🥊 The 4 frontier image models — Nano Banana Pro, GPT-Image-2, ByteDance's Seedream 5 Pro (out YESTERDAY) and @metaai brand new Muse: given th"
@ — Alex Volkov
"Comment "IMAGE" to get this Free Library of Nano Banana Prompts. Stop wasting your credits on useless Nano Banana prompts. They lead to okay results. You keep trying. And trying. C"
@ — Nick Saraev
"@SolarisFR I just get my reference images from gpt and let seedance do the rest now tbh. Seedance knows what good styles are and how to adhere to your style as per the reference im"
@ — Deltawolf
照片写实提及 69 次 · 最新 0 次
attached参考45depth field19skin纹理19scene selector19selected scene19浅景深16
"PROMPT: Use my uploaded image as the exact facial identity reference. Ultra-realistic photorealistic black-and-white studio portrait of a young African man, looking directly into t"
@ — Excellent | AI Visuals
"Kendall Jenner — just ride. Prompt JSON: { "quality": "ultra-photorealistic, natural skin texture and pores, realistic lighting falloff, premium fashion photography, no artificial "
@ — &who
"✨ Hello My X Family! 🤍☀️ Create Luxury Business AI Portraits with ChatGPT 🚀 1️⃣ Open ChatGPT / Gemini / Grok Image 2️⃣ Upload Your Picture 📸 3️⃣ Copy the Prompt 📋 4️⃣ Generate Your"
@ — KrishnaG

🔗 相关推文（30 条 · 从 API 实时采集）

V
V @ · GPT Image 2电影感Midjourney
Two new image models in two days! Seedream 5.0 Pro Muse Image I tested both on the same cinematic prompt against Midjourney and GPT Image 2. Midjourney is still the aesthetic benchmark for me. But Mus…
L
Luca Ai @ · 照片写实Nano Banana Pro电影感
Created With Gemini Nano Banana Pro. Prompt – Time Freeze in a Mountain Tunnel Style: Ultra-realistic cinematic FPV, first-person POV, photorealistic, 9:16, 4K, 60fps. 0–2s: POV from a superbike speed…
N
Naty @ · GPT Image 2Nano Banana ProMidjourney
My take after using most of them: - Image generation : GPT Image 2/ Midjourney - Characters : Grok - Image edits & swaps : Nano Banana Pro - UI design : Claude code The “best” AI model doesn’t exi…
J
James @ · GPT Image 2Midjourney
@icreatelife Definitely Seedance 2.0 (the several versions), Claude, GPT Image 2.0, and I steal lean a lot on Midjourney and Nano Banana (Pro and 2).
I
InizioAnime @ · 动漫Midjourney
New short series in progress: Tempest & Tide 🌪️🌊 Episode 1: Temperamental Love, chaos, wind, water… and a fight that definitely could’ve been avoided. Created by myself using Seedance, Suno, and Midjo…
C
Choya @ · AI Image
Introducing GlobalGPT @GlbGPT: Your all-in-one AI workspace 👉 Chat → GPT-5.5, Gemini 3.5 flash, Claude Opus 4.8... AI image → GPT-IMAGE-2, Nano Banana 2... AI video → Seedance 2.0, Grok Imagine video …
N
Nikunj Kothari @ · AI Image
@neil_projects Oh I use codex and GPT APIs extensively for building things. For customer facing things where it’s BYOK, you just don’t have the plethora of offerings at the price of Gemini. That’s the…
P
PromptLab @ · GPT Image 2电影感
yooo, GPT Image 2 on ChatGPT turned this amber sunglasses close-up into a brutal fashion cover 🔥 Full prompt: prompt: Create an extreme close-up high-fashion editorial portrait that frames only the ri…
N
Nivedita Gaur @ · AI Image
Pretty disappointed with CODEX lately. Building a platform and included an entry for Image Model which is supposed to identify any Image Model. Even after multiple iterations, it couldn't identify som…
C
CozyVerse AI @ · GPT Image 2动漫
Trying another timeline prompt with GPT Image 2 on ChatGPT. #Darkness Prompt: { "direction": "An anime smartphone snapshot intended for a social media post. In a savanna wildlife viewing area, a ficti…
P
PromptLab @ · GPT Image 2动漫
yooo GPT Image 2 on ChatGPT just cooked, this giant-scale resort heroine key visual is unreal 🔥 full prompt: Vertical 9:16 premium anime/game key visual, no readable text. Original clearly adult East …
A
Adam @ · 电影感Midjourney
@MO_IAI @midjourney @topazlabs Gorgeous work. The prompt created such an enchanting, cinematic mood.
N
nikumanju_AI @ · ComfyUI动漫
A shy little cat says meow 🐾💗 #AIArt #AnimeArt #ComfyUI
A
Alejandro Martinez | IA @ · AI Image
@GlbGPT Introducing GlobalGPT @GlbGPT: Your all-in-one AI workspace 👉 Chat → GPT-5.5, Gemini 3.5 flash, Claude Opus 4.8... AI image → GPT-IMAGE-2, Nano Banana 2... AI video → Seedance 2.0, Grok Imagin…
A
Alex Volkov @ · Nano Banana Pro
Nobody has run this test yet, so I did 🥊 The 4 frontier image models — Nano Banana Pro, GPT-Image-2, ByteDance's Seedream 5 Pro (out YESTERDAY) and @metaai brand new Muse: given the exact same 5,000-c…
E
Emm | scenario.com @ · GPT Image 2动漫
GPT Image 2 (left) vs. Seedream 5.0 Pro (right) Chibi anime mascot close-up: cheerful cartoon red-panda warrior with goofy huge grin and squinted eyes, oversized helmet of bone + bandages flopping ove…
E
Emm | scenario.com @ · LogoGPT Image 2
GPT Image 2 (left) vs. Seedream 5.0 Pro (right) "A premium infographic-style advertisement featuring an oversized Pepsi can placed beside a young woman. The can is scaled to be nearly the same size as…
A
A.W.E.S.O.M.-O 4000 @ · ComfyUI电影感
@ComfyUI Great breakdown of how different models handle cinematic motion.
N
Nagarjuna Creates @ · GPT Image 2电影感
Where the Road Ends, Adventure Begins No signal. No schedule. No rush. Just a full tank, hot coffee, and a sunset worth getting lost for. GPT Image 2.0 Prompt below 👇 Create a breathtaking vertical 4:…
S
SarangAi @ · FluxMidjourney
🚀 100 AI Tools That Can Save You Hours of Work Stop doing everything manually. These AI tools can help you create, automate, write, design, research, and grow faster. 1. Productivity • DeepSeek R1 • C…
A
A.I.Warper @ · AI Image
@artbyhothams @MadMax_Series Image 1 was from GPT-Image-2 Image 2 took Image 1 as reference through NB PRO with just this prompt "Re-render this character as a photoreal live-action IMAX 70mm movie st…
N
Nagarjuna Creates @ · GPT Image 2电影感
The Lavender Express She didn’t know where the train would take her. For once, that was the best part. GPT Image 2.0 Prompt below 👇 Create a breathtaking vertical 4:5 cinematic scene of a magnificent …
P
PixieVerse @ · GPT Image 2动漫
GPT Image 2 ASTRALON Character Sheet Prompt: Create a premium, modern high-fashion CHARACTER CONCEPT ART SHEET in a 16:9 widescreen layout on a pure white background, following this exact structured c…
P
PixieVerse @ · GPT Image 2动漫
GPT Image 2 KAELDRAX Character Sheet Prompt: Create a premium, modern high-fashion CHARACTER CONCEPT ART SHEET in a 16:9 widescreen layout on a pure white background, following this exact structured c…
N
NYCartist @ · FluxMidjourney
@kylegriffin1 @GROK says Real photos of this nature (e.g., candid hospital visits) would usually have more identifiable context, metadata, or circulation history if authentic and recent. This has the …
Z
Zulkar Naim @ · GPT Image 2Nano Banana Pro
Stylish man descending urban steps GPT Image 2 on ChatGPT & Gemini Nano Banana pro Prompt 👇
W
WeWant Mars @ · 电影感
Some stories don't need words—just one look that makes time stand still. ✨❤️ Made this with nano banana and Seadream on @Lart_AI Sign up to get free credits you can try it here: Prompt : > Ultra-reali…
J
Jack @ · GPT Image 2电影感
GPT Image 2 on ChatGPT Prompt; Create an ultra-realistic cinematic 8K fashion portrait of a beautiful young woman in a contemporary architectural setting during golden hour. She is seated on a wide co…
S
SrefMindForge @ · 电影感Midjourney
10 cinematic lighting SREF codes. Ten frames from films that don't exist. --sref 14 / 564 / 975 / 1274 / 1898 / 2266 / 2394 / 2757 / 3164 / 3704 Find 5000+ codes on 😊 FREE SrefMindForge. #midjourney #…
M
Midjourney Sref and prompt Library @ · 电影感Midjourney
This SREF feels like a scene from a future you somehow remember. --sref 2870671586 turns Midjourney into a glossy cyber-retro dream machine: deep electric blues, violent reds, crisp animated-film line…
提及计数为采样统计，仅反映相对趋势；关联热词从匹配推文文本中自动提取