Hello and welcome to our weekly roundup!
So, the world has finally seen the long-awaited release of GPT 4.5 and... Not very happy with it? Judging by the first tests and the reaction I see on X, this may be the weakest OpenAI release yet.
But let's get everything in order.
This Creators’ AI Edition:
Featured Materials 🎟️
News of the week 🌍
Useful tools ⚒️
Weekly Guides 📕
AI Meme of the Week 🤡
AI Tweet of the Week 🐦
(Bonus) Materials 🎁
Your advertisement could be featured here!
Sponsor a spot in our newsletter to connect with subscribers interested in AI, technology, and startups. For inquiries, click the button below.
Featured Material 🎟️
OpenAI Unveils GPT-4.5 Preview
OpenAI has launched GPT-4.5, its latest AI model, marking a notable update to ChatGPT. Today, the model was launched as a research preview for ChatGPT Pro users. Next week, it will expand to Plus and Team subscribers. The company describes it as its largest model yet.
GPT-4.5 was trained with more computing power and data than any previous release, aiming to refine how AI interacts with humans.
Key highlights include improved emotional intelligence, which allows it to pick up on subtle cues and respond more empathetically. It also reduces hallucinations and excels at connecting ideas. So, it should be a more potent tool for problem-solving.
Sam Altman called it the "first model that feels like talking to a thoughtful person." In his opinion, chatbots are now moving towards more intuitive conversations. However, it’s not a complete leap to GPT-5. You can think of it as a bridge, with the upcoming GPT-5 to blend this tech with reasoning models like o3.
You can compare this model with others on the OpenAI website.
Keep your mailbox updated with key knowledge & news from the AI industry
Ambiguous Reception
Not everyone is happy with GPT-4.5.
The first thing that grabbed my attention was that the Explore page on X changed from “GPT-4.5 Release” to “GPT-4.5: A Leap in AI or a Step Back?” Many users began to write negative comments based on their impressions. They believe the difference between GPT-4o and GPT-4.5 is not that big.
But things got worse from there. Because the community seems to be right.
Andrej Karpathy (Co-founder and former Chief Scientist of OpenAI) made an interactive comparison of two models, the new one and its predecessor. As blind testing with users showed, in 4 out of 5 cases, people preferred the responses from the old GPT-4o. He also noted that the new model (because of its size) is much slower.
That is, it loses both quality and speed.
Right now, I am not prepared to share my opinion on this release.
In the coming days, we at Creators' AI will be actively testing GPT-4.5 and will get back to you with our verdict very soon. So stay tuned!
News Of The Week 🌍
Perplexity Develops 'Comet' Browser for Agentic Search
Perplexity AI has announced Comet, a new web browser designed to enhance agentic search capabilities. Comet promises to provide users with a more interactive and intuitive experience. We don't have any other details; the company has yet to share the features or look of its next product. However, you can already join the waiting list.
Funnily enough, Perplexity CEO Aravind Srinivas asked users on X what features they would like to see in the new browser. It makes me wonder if the startup isn't sure of the result. Nevertheless, I'm looking forward to trying it out!
Amazon Launches Alexa+
Amazon has unveiled Alexa+, its next-generation assistant based on Bedrock's LLMs. The upgraded Alexa features free-flowing conversations and integrates with over 20,000 services and devices. According to the company, this model also introduces proactive suggestions and cross-device continuity via a new mobile app and browser interface. Priced at $19.99/mo, it will be free for Amazon Prime members.
In its press release, Amazon forgot to mention a key fact: the new Alexa uses Anthropic's Claude as a major model. The partners have worked on the assistant for the past year.
Anthropic Debuts Claude 3.7 with Hybrid Reasoning
Speaking of Anthropic. Earlier this week, the startup released Claude 3.7 Sonnet, its first hybrid reasoning model capable of instant responses or extended, visible problem-solving. The model excels in coding and front-end, with a new Claude Code tool enabling terminal-based agentic coding. Claude 3.7 is available across all Claude tiers (including free) and cloud platforms like Amazon Bedrock.
All the big AI players have immersed themselves in “reasoning.”
Sharing is caring! Refer someone who recently started a learning journey in AI. Make them more productive and earn rewards!
OpenAI Expands Deep Research Access to ChatGPT Plus
OpenAI has democratized its Deep Research tool. It is now available to all ChatGPT Plus, Team, Edu, and Enterprise users for $20/month—previously exclusive to Pro subscribers at $200/month. The feature enables multi-step analysis, synthesizing reports from web sources and user-uploaded files (text, images, PDFs) using the o3 reasoning model.
You'll find everything you need to know about deep search in this post:
ElevenLabs Unveils 'Scribe' Speech-to-Text Model
ElevenLabs has launched Scribe, its latest speech-to-text model. Designed to handle real-world audio challenges, Scribe offers transcriptions across 99 languages. Benchmark tests show it outperforms rivals like OpenAI’s Whisper and Gemini 2.0 Flash, achieving 96.7% accuracy in English and 98.7% in Italian. A low-latency version for real-time use is in development. The model is accessible via API or dashboard uploads, targeting applications like meeting summaries and subtitles.
Alibaba Open-Sources Video-Generating AI Models
Alibaba released four variants of its Wan2.1 series AI models for free public use, enabling text-to-video and image-to-video generation. The open-source models—ranging from 1.3B to 14B parameters—produce short videos up to 720p resolution and beat rivals like OpenAI’s Sora in motion accuracy and visual quality benchmarks.
Available on Hugging Face, they target commercial and academic users.
Useful Tools ⚒️
Helix – From idea to investor-ready prototype in 3 mins
pikr – Receive your summarized Newsletters in Notion
Lemni – Set up custom AI agents in minutes
Pinch – Immersive real-time voice translation for video conferencing
Lex Page — Simple writing tool with AI Editor
Lex Page is a new platform that combines the functionality of a word processor with AI. Its goal is to offer an alternative to Google Docs (and other word processors) with a simple design but many smart features. Using GPT-4, Lex Page generates relevant recommendations for editing and improving text. You can think of it as a personal editor and proofreader that you can call (or silence) when needed.
You can learn more about this tool from the creator's video:
Share this post with friends, especially those interested in AI stories!
Weekly Guides 📕
How to Generate Amazing AI Images with Freepik
Build AI Agents Without ANY Code! (AutoGen Studio Guide)
AI Copywriting Guide & Top Tools For Lifecycle Marketers
A Complete Guide to Claude 3.7 with Code Comparison
AI Meme Of The Week 🤡
AI Tweet Of The Week
Yeah, this isn't the best start for an OpenAI product.
You’ll find a full breakdown of Karpathy here.
(Bonus) Materials 🏆
AI Turned Me Into a Content Agency of One (Must read!)
How I use AI as a writing assistant
What Different Types of AI Are There? — Complete Guide
U.S. Workers Are More Worried Than Hopeful About Future AI Use
Don’t Freak Out About Empathic Chatbots. Learn From Them.
Share this edition with your friends!