Hey everyone! Big news from OpenAI – they just made their super-smart models even better, and we’re here to break it down in easy terms for you!
1. GPT-3.5 Turbo: Faster, Smarter, Cheaper!
Let’s talk about the backbone of most of our virtual conversations – GPT-3.5 Turbo. OpenAI has given it a serious upgrade, making it a powerhouse for understanding and generating natural language or code. Besides, the cool new features include sharper instruction-following skills, support for JSON patterns, more predictable results, and optimisations for multitasking.
However, here’s the real kicker – a brand-new 16k version of GPT-3.5 Turbo! This bad boy handles those extra-long inputs and outputs like a boss. Assuming of it as the assistant of text processing, ready to tackle your wordy challenges with finesse.
Oh, and did we mention the price drop? Yes, you heard it right! OpenAI is slashing the prices for GPT-3.5 Turbo inputs by 50% and outputs by 25%. Now you can dive into those lengthy analyses of papers or books without burning a hole in your pocket. The new rates? Just $0.0005 per 1,000 tokens for input and $0.0015 per 1,000 tokens for output. It’s like a holiday gift for all the text-intensive app enthusiasts out there!
2. Meet GPT-3.5 Turbo 0125
OpenAI quietly introduced GPT-3.5 Turbo 0125 – the unsung hero of this update. While not stealing the spotlight like its sibling GPT-3.5 Turbo 0613, this version brings a series of “various improvements.” It may not shout about its achievements, but hey, every hero has its silent victories!
3. GPT-4 Turbo: No More Laziness!
Furthermore, GPT-4 Turbo had a little “laziness” problem, but OpenAI fixed it! Say hello to the new API usage preview model, GPT-4 Turbo 0125! It not only banishes laziness but also amps up performance, especially in code generation tasks. Still in preview mode, but the full-fledged GPT-4 V is expected to hit the stage in the coming months. Get ready for some mind-blowing advancements!
4. Embedding Models: Smarter and Cheaper Text Tricks!
OpenAI is not stopping there. They’ve unleashed the new text-embeddings-3-small and text-embeddings-3-large models, offering a whopping 3072 dimensions of embedding space. What does that mean for you? More semantic information captured, leading to better accuracy in your tasks.
But here’s the kicker – these new models are not just brainy; they’re also budget-friendly. Text-embeddings-3-small is 1/5 the price of its predecessor, text-embeddings-ada-002. Developers, rejoice! You’ve got a cost-effective option that doesn’t compromise on performance.
5. Text Audit API: Your Virtual Content Checker Got an Upgrade!
OpenAI didn’t forget about making sure everything’s safe and sound. They upgraded their free Text Audit tool. Now, it’s better at checking for anything sensitive or not so nice in your words. It’s like having a virtual helper watching over your writing.
In a nutshell, OpenAI’s updates are like giving our virtual pals a big brain boost! They’re faster, smarter, and even friendlier on our wallets. So, whether you’re chatting, creating code, or just having fun with words, OpenAI’s got your back with these awesome upgrades. Get ready for a whole new world of smart text adventures!
