⚡ OpenAI unveils newest o3 model

In partnership with

Hi AI Futurists,

Today, we’re focusing on OpenAI’s latest breakthrough: the o3 model, a powerful new system redefining AI’s capabilities in reasoning and problem-solving. We’ll explore its record-breaking performance, the upcoming cost-effective mini version, and what it all means for the future of technology, creativity, and productivity.

Here’s our agenda.

• Fellow

• o3, OpenAI’s latest and most advanced model

• Top 3 selected AI tools

• Top news on the AI horizon

• Writer

Best,
Lex

Automate your meeting notes

Get the most accurate and secure meeting transcripts, summaries, and action items.

Never take meeting notes again: Fellow auto-joins your Zoom, Google Meet, and MS Teams meetings to automatically take notes.

Condense 1-hour meetings into one-page recaps: See highlights like action items, decisions, and key topics discussed.

Claim 90 days of unlimited AI notes today.

Get started free

How AI is Impacting the World

OpenAI’s o3 model is coming soon

OpenAI’s upcoming o3 model marks a breakthrough in AI, achieving unparalleled performance in complex reasoning tasks. Building on its predecessor, o1, o3 sets new standards in benchmarks like the ARC Prize, hitting 87.5% accuracy with enhanced computing resources. This is a major leap toward artificial general intelligence (AGI). François Chollet, creator of the ARC benchmark, describes o3 as an "important step-function increase," likening its methodology to AlphaZero’s exhaustive problem-solving approach, though at a high computational cost.

The o3 model’s prowess extends beyond benchmarks. In the challenging Frontier Math Benchmark, it achieved 25.2% success, far exceeding previous models’ sub-2% performance. In practical applications like programming and PhD-level science, o3 has surpassed human experts, earning a Codeforces score of 2727 and scoring 87.7% on science questions. This underscores o3’s ability to generate and execute real-time solutions, moving beyond traditional pattern-matching models. However, such feats require intensive processing—up to 33 million tokens for a single task—driving costs as high as $20 per task.

While o3 isn’t AGI yet, its mini version, launching in January 2025, aims to balance performance and affordability. Demonstrations showcased its ability to autonomously generate Python scripts and manage structured outputs, outperforming o1 even at lower settings. As AI capabilities surge, the line between human and machine problem-solving continues to blur, raising profound questions about the role of intelligence in human creativity and productivity.

Important Points

• o3 achieves 87.5% on the ARC benchmark, significantly surpassing its predecessor and even human experts in certain areas.

• Tasks cost up to $20 each due to the system’s intensive computing, processing as many as 33 million tokens per task.

• A mini version is set to launch in January 2025, combining scalable performance with lower costs and surpassing o1’s capabilities.

What I’ll be doing

I believe o3 is a game-changer, solving complex problems and outperforming humans in key areas. High costs are a hurdle, but the upcoming mini version makes advanced AI more accessible.

I’ll be staying tuned for o3 Mini’s launch to explore how its capabilities can streamline my workflows. I’ll also be testing o3’s tools and digging into its potential for automation and creativity.

Apply AI Superpowers with Tools

1. Websparks

   The AI software engineer that brings your ideas to life

   

2. Nexal AI

   Text-to-speech AI voiceovers in more than 140 languages

   

3. GrantGPT

   Your public funding matchmaker: Research & apply in minutes

   

On the Horizon

• Arizona’s getting an online charter school taught entirely by AI

• Google Search will reportedly have a dedicated ‘AI Mode’ soon

• Apptronik Partners with Google DeepMind

• Ex-Twitch CEO Emmett Shear is founding an AI startup backed by a16z

• OpenAI rolls out enhanced memory for ChatGPT, allowing it to reference previous chats

• Gemini can now tell when a PDF is on your phone screen

• Apple might be working on a smart doorbell

• OpenAI’s GPT-5 reportedly falling short of expectations

• 10 AI Predictions For 2025

• Instagram teases AI editing tools that will completely reimagine your videos

Writer RAG tool: build production-ready RAG apps in minutes

RAG in just a few lines of code? We’ve launched a predefined RAG tool on our developer platform, making it easy to bring your data into a Knowledge Graph and interact with it with AI. With a single API call, writer LLMs will intelligently call the RAG tool to chat with your data.

Integrated into Writer’s full-stack platform, it eliminates the need for complex vendor RAG setups, making it quick to build scalable, highly accurate AI workflows just by passing a graph ID of your data as a parameter to your RAG tool.

Learn more about our production ready RAG tooling here.

That’s all for today, folks!

• If you’re enjoying the newsletter, share with a friend by sending them this link: 👉 https://www.futureblueprint.xyz/subscribe

• Looking for past newsletters? You can find them all here.

• Working on a cool A.I. project that you would like us to write about? Reply to this email with details, we’d love to hear from you!