Daily Digest: Mini models are insane

PLUS: GPT-4o-mini and Mistral NeMo are latest additions.

Want to get in front of 100k AI enthusiasts? Work with us here

Hello folks, ICYMI - we:

Here’s what we have today;

PICKS

  1. New tutorials:

  2. OpenAI has a new baby - GPT-4o-mini. It is a small model that beats the early version of GPT-4 and it costs 30x less than its big bro GPT-4o. You can now say bye to GPT-3.5-Turbo as 4o-mini will take its place in ChatGPT’s free tier and the API.🍿Our Summary (also below)

    ps: New voice mode in ChatGPT is also coming soon.

  3. Mistral AI, in cahoots with NVIDIA, released a new LLM - Mistral NeMo. It is kinda big (12B parameters) compared to its peers (Gemma 2 9B and Llama 3 8B). But it packs a 128k context window and some impressive multilingual skills. Oh, and it's open-source.🍿Our Summary (also below)

  4. The fastest way to build AI apps

    • Writer Framework: build Python apps with drag-and-drop UI

    • API and SDKs to integrate into your codebase

    • Intuitive no-code tools for business users

    Start building with AI Studio

TOP TOOLS

  • OnDemand* - Unlock exclusive access to curated AI solutions (with a seamless, user-friendly interface). Join now and transform your projects.

  • Flow Studio - Fully-produced, film-quality 3-minute videos with stories, characters, and sound, all in one click.

  • Workflow Creation in Julius - Create a repeatable, templated AI workflow for data analysis and offload it to Julius.

  • Proton Scribe - A private writing assistant that writes and proofreads emails for you.

  • Archie - Design and plan software applications in minutes with AI.

  • Storytime - Turn family photos into beautifully illustrated children’s bedtime stories.

  • Buildco - Build MVPs with Next.js using AI in minutes.

View more →
*sponsored

NEWS
QUICK BITES

OpenAI has a new baby - GPT-4o-mini. It is a small model that beats the early version of GPT-4 and it costs 30x less than its big bro GPT-4o.

What is going on here?

OpenAI's releasing GPT-4o mini, a small but mighty AI model that's way more affordable than its big brothers.

What does this mean?

GPT-4o mini is a powerhouse in a small package, designed to replace GPT-3.5 models. It's incredibly affordable at just 15 and 30 cents for a million input-output tokens, making it 60% cheaper than GPT-3.5 Turbo and ~30x budget-friendly than top-tier models.

In fact, it even beats out other small models like Google's Gemini 1.5 Flash ($0.35/$0.70) and Anthropic's Claude 3 Haiku ($0.25/$1.25) on price.

But it's not just about the cost savings; GPT-4o mini is seriously smart. It outperforms other small models in math, coding, and multimodal reasoning. , On the MMLU benchmark (general intelligence), it scores 82%, surpassing GPT-3.5 and some larger models.

This little model can handle a massive 128K token context window and outputs 16k tokens, opening up a ton of new possibilities. Companies like Ramp and Superhuman are already using it to great success in real-world tasks.

Plus, it's multimodal, just like its bigger sibling GPT-4o, supporting both text and vision inputs with even more on the horizon.

Safety is taken care of too. OpenAI has baked in new techniques like "instruction hierarchy" to keep the model secure and resistant to jailbreaks.

Why should I care?

You can now say bye to GPT-3.5-Turbo as 4o-mini will take its place in ChatGPT’s free tier. If you were using GPT-3.5-Turbo API in your applications, you should switch to 4o-mini.

That leads to two reasons to care: People are getting increasingly smarter AIs for free directly via ChatGPT and app developers can build powerful AI tools without breaking the bank.

This model (and other kiddos like Gemini 1.5 Flash and Calude 3 Haiku) are great for low-logic tasks. Think translations, rewrites, getting data from forms/images etc. Just don’t expect them to use their own brain, and you’ll be fine.

QUICK BITES

Mistral's latest AI model, NeMo, is here to shake things up. It's small but mighty, packing a 128k context window and some impressive multilingual skills. Oh, and it's open-source. Time to get coding!

What's going on here?

Mistral AI, in cahoots with NVIDIA, released a new open-source language model called Mistral NeMo.

What does this mean?

Mistral NeMo is kinda big (12B parameters) compared to its peers (Gemma 2 9B and Llama 3 8B). But Mistral thinks that local machines can still run it and get actual stuff done, not just use open-source as a play toy.

  • It's got a massive 128k token context window (that's a lot of room for chat)

  • Performs like a champ on reasoning, knowledge, and coding tasks

  • Speaks a bunch of languages fluently (not just English, folks)

  • Uses a fancy new tokenizer called Tekken that's super efficient with different languages and code

  • Comes in both pre-trained and instruction-tuned flavors

  • Licensed under Apache 2.0, so it's free for research and commercial use

Oh, and it's quantization-aware, meaning you can run it in FP8 without losing performance. Nerdy, but cool. It’s available at the usual places like HuggingFace and Mistral’s La Plateforme, as well as as a package on Nvidia’s NIM microservice.

Why should I care?

If you're into AI (and who isn't these days?), this is big news. Mistral NeMo brings near top-tier performance in a smaller, more efficient package. This means:

  • Easier and cheaper to run for smaller companies and researchers

  • Better multilingual support for global applications

  • Potential for more diverse and creative AI applications due to its open-source nature.

For developers, it's a drop-in replacement for Mistral 7B, so upgrading should be a breeze. And for the open-source AI community, it's another step towards democratizing powerful language models. Time to play with some new toys!

Ben’s Bites Insights

We have 2 databases that are updated daily which you can access by sharing Ben’s Bites using the link below;

  • All 10k+ links we’ve covered, easily filterable (1 referral)

  • 6k+ AI company funding rounds from Jan 2022, including investors, amounts, stage etc (3 referrals)

Reply

or to participate.