Video wird geladen...
Video konnte nicht geladen werden
Gemini 4 Argon is the first model where It felt like an image, video and language model all at once. Everything in this thread is made with code. Google AI Studio More examples + the full code for all 534 paintings below ⬇️
81,785 Aufrufe • vor 7 Stunden •via X (Twitter)
21 Kommentare

@GoogleAIStudio such cool examples

@GoogleAIStudio Incredible model Logan! Congratulations to the team 🌸

1. Palm springs lawn chair poolside 2. Nacre accordion keys 3. Fata morgana tanker truck 4. Holographic foil marathon cape

1. Shin hanga snowy bicycle crossing 2. Aquatint empty pool steps 3. Fluorescence moth head 4. Swiss grid jacket zip

1. Whiteout hydrant fragment 2. Vorticist jackhammer asphalt 3. Suitcase salt flat horizon 4. Copperplate cracked walnut

all 534 paintings, and the full code behind every one, are here:

@GoogleAIStudio Wow! Love it!

@GoogleAIStudio Thank you and congratulations!

@GoogleAIStudio This is probably the coolest thing ever

@GoogleAIStudio @kickingkeys couldn't have said it better! ty ser

@GoogleAIStudio I love these.

@GoogleAIStudio you're cooking so hard it's unbelievable

@GoogleAIStudio thank youu Pauline!!

@GoogleAIStudio these are incredible!

@GoogleAIStudio impressive work

@GoogleAIStudio Wow, awesome stuff!

@GoogleAIStudio Giving off unmistakable Opus and Fable vibes here, or is that just me?

@GoogleAIStudio 534 paintings says the 'felt like' is more than just a feeling.

@GoogleAIStudio 🔥

@GoogleAIStudio these are so cool

@GoogleAIStudio how does it handle ambiguous diagrams where the text labels matter more than the shapes? that's where most vision models still break.
