As February came to its final day, the last week of the month was nothing but filled with AI updates. OpenAI unveiled its latest GPT-4.5 model, Microsoft announced its new Phi-4 small language models (SLMs), Google revealed pricing for its Veo 2 video generation tool, and a lot more.
So, if you are too overwhelmed with all the latest happenings in the AI space, then you have arrived at the right place. In this story, I have compiled all the latest AI happenings from the past week. So, without wasting any more time, let’s jump right in:
OpenAI Introduces GPT-4.5
OpenAI has released its latest and most advanced AI model called GPT-4.5. This competes with the likes of recently launched Grok 3 and DeepSeek V2. The new version is said to bring improvements in reasoning, speed, and efficiency, marking another step in the company’s development of large language models.
Currently, the GPT-4.5 model is only available for ChatGPT Pro users and developers. However, the company is planning to extend the availability to Plus and Team users in the coming weeks, followed by Enterprise and Education users.
Microsoft Introduces Phi-4 AI Models
Microsoft also took to the stage council its new AI models called Phi-4-Multimodal and Phi-4-Mini. These are small language models (SLMs) and are part of Microsoft’s Phi family. It claims to enhance AI capabilities without compromising on efficiency and scalability. Both models are available on Azure AI Foundry, Hugging Face, and the NVIDIA API Catalog.
Google Challenges GitHub Copilot with Free AI Coding Tool
We’re launching a free version of Gemini Code Assist globally to help you build faster. It comes with:
— Google DeepMind (@GoogleDeepMind) February 25, 2025
🛠️ 180K code completions per month
🌐 Support for all programming languages in the public domain
💡 128K token context window
Get started → https://t.co/w12E5VDPI3 pic.twitter.com/UJv98rre8X
Google also launched its free version of the Gemini Code Assist. Earlier, it was only available to enterprise users. But now, the tech giant has unveiled its free variant globally for individual developers, freelancers, and more.
It’s built on Google’s advanced Gemini 2 AI model and is capable of writing codes faster and more efficiently. The tool also gives real-time code completions, explanations, and AI-powered suggestions, making it easier for programmers to streamline their workflow.
Google Unveils Veo 2 Pricing: Costs 50 Cents Per Second
Apart from the AI coding assistant, Google also unveiled the pricing details of its latest video generation model called Veo 2. It was launched in December last year and competes against the likes of OpenAI’s Sora. However, this tool will cost 50 cents per second. This translates to $30 per minute and a substantial $1,800 per hour for video generation.
Paytm Integrates AI-Powered Search with Perplexity
India’s payments pioneer is here with some exciting news! Paytm x @Perplexity_AI are coming together for an insightful session on AI, mobile payments, and more.
— Paytm (@Paytm) February 27, 2025
Tune in live as we showcase how AI is enhancing the experience for millions of consumers #PaytmKaro https://t.co/tjpyyJqJXO
Paytm has partnered with AI-powered search platform Perplexity to integrate a search interface into its apps. The new search experience, available as a free tool, allows users to get real-time answers to queries and financial assistance without leaving the app. It is available in the “Free Tools” section and features a full-screen user interface with basic features from Perplexity.
DeepSeek Launches 75% Off AI Model Sale During Off-Peak Hours
🚨 Off-Peak Discounts Alert!
— DeepSeek (@deepseek_ai) February 26, 2025
Starting today, enjoy off-peak discounts on the DeepSeek API Platform from 16:30–00:30 UTC daily:
🔹 DeepSeek-V3 at 50% off
🔹 DeepSeek-R1 at a massive 75% off
Maximize your resources smarter — save more during these high-value hours! pic.twitter.com/00Qa6iNbcG
DeepSeek has announced a massive discount program, offering up to 75% off its AI models during off-peak hours. From 16:30 to 00:30 UTC daily, users can access DeepSeek-V3 at a 50% discount, while DeepSeek-R1 is available at a whopping 75% off. This move aims to encourage users to optimize resource usage during less busy periods, potentially disrupting the AI market and challenging competitors in the US and Europe. The discounted pricing applies to both input and output tokens, making it an attractive option for developers seeking cost-effective AI solutions.
That’s a wrap with the AI updates for now. But if you are interested in staying up to date with all the latest happenings in the AI and business tech space, then make sure to follow Geekflare on Google News. We’ll keep you posted!
