In partnership with

 🦾 GPT-4.1: OpenAI’s Newest Brainiac with a Million-Token Appetite.

Welp… remember yesterday when I said a reverse engineer spotted OpenAI’s plans to drop new models soon — but we didn’t know when?

Yeah. About that.

OpenAI pulled a fast one and casually launched not one, not two, but three new models yesterday — GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano.

They’re here. They’re smart. They’re built for devs who are tired of living on a diet of Stack Overflow and vibes.

These models especially the headliner, GPT-4.1, is designed to:

• Crush coding tasks like a pro (especially frontend work)

• Actually follow instructions without going rogue

• Stick to formats, use tools correctly, and not randomly rewrite things just because it feels like it

• Handle up to 1 million tokens in one go — that’s around 750,000 words (aka longer than War and Peace, but hopefully less dramatic).

But here’s the catch: you won’t find it in ChatGPT (yet). These models are API-only for now.

Now let’s talk speed vs. smarts.

While GPT-4.1 Mini is cheaper, faster, and still pretty clever, GPT-4.1 Nano is even cheaper, even faster—but it trades off some accuracy.

Basically, Nano is your speed demon — great for quick tasks, but maybe don’t ask it to build your next app from scratch.

And the pricing? Buckle up:

• GPT-4.1 costs $2 per million input tokens, and $8 per million output tokens

• GPT-4.1 Mini: $0.40 in, and $1.60 out

• GPT-4.1 Nano: $0.10 in, and $0.40 out (aka your wallet’s new best friend)

But how smart is it really?

Well…

• On the SWE-bench coding benchmark, GPT-4.1 scored between 52% and 54.6%

• That’s solid — but not chart-topping. For context, Google’s Gemini 2.5 hit 63.8%, and Claude 3.7 from Anthropic pulled 62.3% 

• However, GPT-4.1 did flex on video comprehension, scoring 72% accuracy on long videos with no subtitles — so yeah, it might just out-binge-watch you.

But let’s not ignore the fine print.

OpenAI was pretty upfront about the flaws. For starters:

• The more tokens you give it, the more its accuracy tanks — as OpenAI says, the accuracy dips from 84% at 8K tokens to 50% at 1 million

• It can also be a little too... literal. If you’re not specific, it might follow your prompt like a passive-aggressive intern doing exactly what you said, but not what you meant.

Bottom line?

OpenAI clearly wants to build a full-on AI software engineer — something that can code, debug, QA, and even write the documentation. And GPT-4.1 is a solid step in that direction.

But before we break out the confetti and hand them the “Best AI Ever” award, it’s worth asking: Will it live up to the hype?

(Let’s not forget — Meta already reminded us how sketchy benchmarks can be.)

In the end, we’ll wait to see what real devs say once they’ve had time to test these models in the wild. Until then? We’re holding off on the standing ovation.

Full details here.

Hire Ava, the Industry-Leading AI BDR

Your BDR team is wasting time on things AI can automate. Our AI BDR Ava automates your entire outbound demand generation so you can get leads delivered to your inbox on autopilot.

She operates within the Artisan platform, which consolidates every tool you need for outbound:

• 300M+ High-Quality B2B Prospects, including E-Commerce and Local Business Leads

• Automated Lead Enrichment With 10+ Data Sources

• Full Email Deliverability Management

• Multi-Channel Outreach Across Email & LinkedIn

• Human-Level Personalization

Book a demo to see what Ava can do.

 💬 Meta's AI Revolution: Why Your Posts Are More Valuable Than Ever

So, Meta’s making moves in Europe again, but this time, it’s all about AI.

They’ve just announced they’re going full steam ahead with training their AI models using public content from Facebook and Instagram users in the EU—we’re talking posts, comments, interactions—basically, everything you see on your feed.

But don’t worry, they’re not touching private messages or anything from users under 18, so no need to panic.

Now, you might be wondering, why now? Well, last year, the privacy storm hit hard, and Meta had to pull back.

Since then, they’ve gotten clearer guidance from regulators, so they’re back on track and ready to launch this week.

For you, the user, that means you’ll get notifications both in-app and by email—letting you know that your public posts might be used to train AI.

And of course, you can opt out if you’re not cool with that.

Meta’s also making sure they honor all opt-out forms that have already been submitted, along with any new ones that come in.

So, why does all this matter?

Well, Meta’s goal is to create AI that truly understands Europe.

And by “understands,” they mean getting the local dialects, the quirky humor, and those cheeky sarcastic comments that make the internet such a fun place.

Basically, they want their AI to really “get” the different vibes across Europe, so it can respond in a way that feels more... real. Makes sense, right?

And while Meta’s not the only one using European data—Google and OpenAI are already deep into it—Meta’s adding its own European twist to the whole thing.

And don’t think the regulators are just chilling on this. The Irish Data Protection Commission (DPC) is still keeping a watchful eye. Just last week, they announced they’re investigating xAI’s Grok AI model and how it’s being trained.

After all, AI is still on a pretty tight leash in Europe.

The Bottom line?

Meta’s moving forward with AI training, but they’re playing by the European rulebook.

So, if you’re okay with it, your public posts will help build smarter AI models. If not, no sweat—you’ve got the option to opt out.

Either way, AI in Europe is about to get a little more human, a little more local, and way more tuned in to what makes Europe Europe.

Full details here.

Free Unlimited AI, Need we say more?

Running a startup is complex. That's why thousands of startups trust Notion as their connected workspace for managing projects, tracking fundraising, and team collaboration. 

Apply now to get up to 6 months of Notion with unlimited AI free ($6,000+ value) to build and scale your company with one tool. 

Try FREE Now.

🧱Around The AI Block

• 💼 Some of the most interesting startups showcased at Google Cloud Next.

• 👨‍💻 How to use Claude to create a virtual assistant, in under 10 minutes.

• ❌ OpenAI plans to phase out GPT-4.5, its largest-ever AI model, from its API.

• 😱 Debates over AI benchmarking have reached Pokémon

• 👨‍🏫 Google Classroom gives teachers an AI feature for quiz questions.

• 🐬 Google’s newest AI model is designed to help study dolphin ‘speech’.

• 💪 Nvidia wants to manufacture some AI chips in the US.

🛠️ Trending Tools

• Meme Alerts uses AI to send daily viral meme templates to help social media managers and creators stay on top of trending content.

• Central automates payroll, benefits, and compliance tasks for startups, saving founders time and hassle.

• Clerkly is a search engine that answers questions about U.S. regulations using only official government sources.

• AI Eraser is a Chrome extension that helps users redact personal information from their ChatGPT prompts, running entirely on-device to protect user privacy.

• MYND provides personalized meditation, 24/7 support through an AI companion, plus various other wellness features to help you manage your mental health.

🤖 ChatGPT Prompt Of The Day: Find Data Sources.

The hunt for relevant, credible data sources can sometimes feel like looking for a needle in a haystack.

But with the right ChatGPT prompt, you can quickly pinpoint the sources that will act as a foundation for your research paper.

See this as an AI WoD!

Act as an academic research expert. Your task is to identify and compile a list of credible data sources related to [topic]. Ensure that the sources are reputable, recent, and relevant to the research objectives. Your list should include both primary and secondary data sources, such as academic journals, government databases, industry reports, surveys, and any other pertinent repositories. Provide a brief description for each source, highlighting its relevance and credibility. Additionally, make note of any access restrictions or subscription requirements for the sources. Aim to provide a comprehensive list that will serve as a strong foundation for the research on [topic].

Source

Premium members get daily video prompts, premium newsletter, an no-ad experience - and more!

Already a paying subscriber? Sign In.

Premium members get::

• • 👨🏻‍🏫 A 30% discount on the AI Education Library (a $600 value - and counting!)
• • 📽️ Get the daily AI WoD (a $29.99 value!)
• • ✅ Priority help with AI Troubleshooter
• • ✅ Thursday premium newsletter
• • ✅ No ad experience
• • ✅ and more....