Gemini 2.5 Pro: Google macht das derzeit stärkste KI-Modell überraschend kostenlos für alle - Innovationen - derStandard.de › Web
https://www.derstandard.de/story/3000000263527/gemini-25-pro-google-macht-das-derzeit-staerkste-ki-modell-ueberraschend-kostenlos-fuer-alle #KI #Gemini #Gemini25Pro

Did the mighty scientific AI coding shootout of Gemini 2.5pro and ChatGPT o1: same prompt to build a sudoku, a backgammon, a go app in python. Gemini beat ChatGPT soundly. What I find most impressive is not the good code it writes, but the thinking process it shows. Unlike eg. DeepSeek when I read how it approaches the programming task that is DAMN CLOSE to how a good developer thinks. Breaking down problems into smaller ones, keeping an eye on special cases etc.

Maybe I should keep a log of the biggest piece of AI-generated code that worked flawlessly.

Yesterday: 679 lines of python, generated by Google Gemini 2.5 pro.

I remember, half a year ago, I was impressed by something like 100 lines of Javascript code that worked without any bug.

Would be an interesting graph to see...

#Gemini25Pro is here! Google's latest AI model focuses on enhanced reasoning, code generation, and multimodal processing.
Ranked #1 on LMArena for human-preferred AI responses
Excels in math, science & logic-based tasks
1M token context window (expanding to 2M soon!)

More on #InfoQ https://bit.ly/41SO8WU

Google stellt KI-Modell Gemini 2.5 Pro überraschend kostenlos zur Verfügung
Google macht sein derzeit stärkstes KI-Modell Gemini 2.5 Pro überraschend ko
https://www.apfeltalk.de/magazin/news/google-stellt-ki-modell-gemini-2-5-pro-ueberraschend-kostenlos-zur-verfuegung/
#News #Services #DeepSeek #Gemini25Pro #GeminiAdvanced #Google #GoogleChatbot #Innovation #KIBranche #KIModell #KIModus #KostenloseVerfgbarkeit #KnstlicheIntelligenz #LLM #MachineLearning #OpenAI #Reasoning #Sprachmodell