Oke, akhir-akhir ini gw lagi suka banget sama DeepSeek.
Terus ketemu CrofAI dan ini DeepSeek V4 Pro termurah yang gw nemuin sejauh ini
→ $0.30 / 1M input
→ $0.50 / 1M output
→ Cache hit cuma $0.003
Bandingin sama
• DeepSeek direct $0.435 input, $0.87 output
• OpenRouter harga sama persis, zero markup
CrofAI masih 30-43% lebih murah.
Bukan cuma harga, speed 104 token per detik, context 1M. Model sama, harga beda jauh.
Ada yang lebih murah lagi dari ini?
JIRRRR!! Nemu library prompt gila lagi, kalian wajib bookmark dan cobain 😍🔥
Lagi-lagi, ini sharing, BUKAN endorse. Gue cobain sendiri copas prompt-nya ke Gemini… dan hasilnya segini cakepnya.
Mata gue melotot, detail hidup, ala foto studio dan dan vibe pas banget sama harapan gue.
Langsung cobain sendiri deh…
We just released Gemma 4 — our most intelligent open models to date.
Built from the same world-class research as Gemini 3, Gemma 4 brings breakthrough intelligence directly to your own hardware for advanced reasoning and agentic workflows.
Released under a commercially permissive Apache 2.0 license so anyone can build powerful AI tools. 🧵↓
Qwen3.6-Plus is live! 🚀
Alibaba's new flagship crushes benchmarks:
🔹 61.6 on Terminal-Bench 2.0
🔹 86.0 on MMMU
🔹 Autonomous: Handles full dev loops from plan to deploy.
The era of Agentic AI is here.
Try it: https://t.co/juSMm2L0VE
So we know Gemma 4 is good at tool calling, but what about web coding?
I threw 4 UI screenshots at three Gemma 4 models and said rebuild this, one shot, no hand-holding, just image in, code out.
Model lineup:
- E4B
- 26B A4B (MoE)
- 31B Dense
(skipped the E2B this round)
Let me know which one you think cooked the hardest
Be sure to update your Grok app, the new Imagine 'Quality' mode is absolutely insane.
It really responds exceptionally well to specific camera terminology and feels far more stylized. I love it! Amazing work from xAI once more :).
Prompt: a gorgeous Kodachrome film still of Santa Barbara, cinematic and hyperrealistic, 2020s, motion blur, anti aliasing, lens distortion
Google Veo 3.1 lets you clone yourself from a single selfie and produce daily content without ever recording a video...
here's the workflow:
STEP 1: analyze your selfie with Gemini
- upload a photo of yourself
- prompt: "analyze this photo and write a detailed Veo 3 character description including facial features, hair, skin tone, clothing, and body language"
- Gemini outputs a character profile
STEP 2: generate video prompts
- take that character description
- ask Gemini: "write a 100-150 word cinematic Veo 3 prompt with character appearance, camera angle, and dialogue in quotes"
some key tricks: start with "a selfie video of...", make the arm visible (sells the selfie illusion), add natural eye movement, put dialogue in quotes (Veo generates synced audio) and specify "no subtitles"
STEP 3: generate with Veo 3.1
- paste into Google AI Studio
- use "Ingredients to Video" : upload 4 reference photos for face consistency
- output: 8-second clips at 1080p with synchronized dialogue
you can create a lot of formats with this...
the one going most viral: selfie-style walking + talking to camera... viewers genuinely can't tell it's AI
Listen up 🔊Gemini 3.1 Flash Live is launching today, making a big difference for developers who are building real-time voice and vision agents.
How, you ask? Well, this model delivers:
— Responses that feel as fast as natural dialogue
— Better task completion in noisy environments
— Improvements in complex-instruction following
Vibe code at the speed of thought with Gemini 3.1 Flash Live. Here’s an example to get you started.
Using the model in @GoogleAIStudio, you can build apps as you talk out loud with a pace that keeps up with your brainstorms. Start creating your own app with voice control today, or remix ours: https://t.co/TYubZborBH
shower time 🚿
Nano Banana Pro
{
"subject": {
"description": "A woman with a warm, tan skin tone and long, straight, jet-black hair parted down the middle, flowing over her right shoulder. She has a refined facial structure with high cheekbones, dark almond-shaped eyes with subtle winged eyeliner, well-defined dark eyebrows, and full, plump lips with a natural pinkish-mauve hue. Her skin texture is hyper-realistic, showing fine pores and a natural glow on the cheeks and nose.",
"anatomy_fidelity": "Anatomically accurate to the reference with specific emphasis on a very full, heavy bust volume that remains the most prominent feature of the torso. The chest mass shows natural outward and forward projection, visibly deeper than the ribcage, with soft tissue affected by gravity. Shoulders are soft and slightly angled, and the neck is slender, transitioning into a collarbone partially obscured by hair and clothing.",
"clothing": "A thick, plush, white terry cloth bathrobe or hotel robe, worn open and slouching off the shoulders to reveal the upper chest and cleavage."
},
"pose": {
"orientation": "Three-quarter view of the body, slightly angled to the viewer's left, while the head is tilted subtly toward the right shoulder.",
"gaze": "Direct, sultry eye contact with the camera lens.",
"posture": "Relaxed, leaning slightly forward, which emphasizes the volume and depth of the chest. The right arm is positioned down by the side, while the left hand is partially visible at the bottom right, lightly clutching the edge of the white robe."
},
"environment": {
"location": "A modern, minimalist interior, likely a luxury hotel room or dressing area.",
"details": "A neutral-toned, beige or light taupe vertical wall panel/pillar is directly behind the subject. To the left, a white wall with a faint light switch or outlet is visible. In the background to the right, a glimpse of a bed with white linens and a dark headboard is seen through a doorway or partition."
},
"camera": {
"shot_type": "Close-up portrait focusing on the head, shoulders, and upper torso.",
"angle": "Slightly high angle, looking down toward the subject to capture depth and volume.",
"lens_characteristics": "Standard focal length (approx. 35-50mm) to avoid wide-angle distortion while maintaining a shallow depth of field that keeps the subject sharp and the background softly blurred."
},
"lighting": {
"type": "Soft, diffused indoor lighting, likely from a large window or softbox positioned to the front-left.",
"effects": "Gentle highlights on the forehead, bridge of the nose, and tops of the cheeks. Soft, natural shadows define the contours of the collarbone and the cleavage. Realistic subsurface scattering on the skin creates a healthy, lifelike warmth."
},
"mood_and_expression": {
"expression": "Calm, confident, and subtly provocative.",
"mood": "Intimate, relaxed, and high-end aesthetic."
},
"style_and_realism": {
"format": "Photorealistic digital photography.",
"texture": "Visible skin grain, individual hair strands, and the tactile fiber texture of the white cotton robe. No artificial smoothing or airbrushing."
},
"colors_and_tone": {
"palette": "Neutral and warm. Dominant colors include the stark white of the robe, the deep black of the hair, and a range of tan, peach, and bronze skin tones against a beige background.",
"contrast": "Medium contrast with a focus on tonal transitions in the skin and fabric folds."
},
"quality_and_technical_details": {
"resolution": "8k UHD, high dynamic range, sharp focus on facial features, realistic shadows, micro-detail in fabric fibers.",
"aspect_ratio_and_output": "3:4"
},
"controlnet": {
"pose_control": {
"model_type": "DWPose",
"purpose": "Exact skeletal and pose lock",
"constraints": [
"preserve shoulder width",
"preserve head tilt",
"preserve specific spine curvature",
"preserve hand placement on robe"
],
"recommended_weight": 1.0
},
"depth_control": {
"model_type": "ZoeDepth",
"purpose": "Depth, volume, and camera-to-body spatial lock",
"constraints": [
"preserve chest foreground dominance",
"prevent flat or compressed depth",
"maintain clear torso-to-background separation",
"lock the 3D volume of the bust projection"
],
"recommended_weight": 0.85
}
},
"negative_prompt": {
"forbidden_elements": [
"anatomy normalization",
"body proportion averaging",
"smaller bust than reference",
"reduced chest volume",
"flattened or compressed breasts",
"tightened, lifted, or artificially supported breasts",
"slimmed torso",
"aesthetic proportion correction",
"beauty standard enforcement",
"dataset-average female anatomy",
"camera angles that reduce volume",
"wide-angle distortion not in reference",
"lens compression not in reference",
"cropping that removes volume",
"depth flattening",
"mirror selfies",
"phone-in-hand selfies",
"reflections",
"beautification filters",
"skin smoothing",
"plastic skin",
"airbrushed texture",
"stylized realism",
"editorial fashion proportions",
"more realistic reinterpretation",
"naturalization of prominent features"
]
}
}
Meet Codex for Students.
We're offering college students in the U.S. and Canada $100 in Codex credits.
Our goal is to support students to learn by building, breaking, and fixing things.
https://t.co/WrOtW8E8Lk