- Veroneus Newsletter
- Posts
- OpenAI reveals the GPT-4o Mini!
OpenAI reveals the GPT-4o Mini!
Plus: New Small AI NeMo
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/a8868bea-5fae-4882-ae43-fe5351945374/Veroneus_Newsletterm1.png?t=1719171489)
Welcome AI-Powered Leaders!
Good morning. In today's newsletter, we will be sharing the latest selected news about the AI world.
In today’s topics:
OpenAI Reveals the GPT-4o Mini
OpenAI Aspires to Develop Its Own AI Chip
Mistral AI and Nvidia just introduced Mistral NeMo
Nvidia is Developing a New AI Chip for the Chinese Market
Read Time: 6 Minutes
LATEST NEWS
OpenAI’s Reveals the GPT-4o Mini
The Veroneus: OpenAI has announced the release of GPT-4o mini as a cost-effective and compact variant of its flagship model, GPT-4o. This is to increase AI's reach to developers and businesses.
The Details:
Scores 82% on the MMLU benchmark, outperforming Google's Gemini Flash (77.9%) and Anthropic's Claude Haiku (73.8%).
15 cents per million input tokens, 60 cents per million output tokens (60% cheaper than GPT-3.5 Turbo).
Supports a 128K token context window, handles text and vision inputs. Audio and video capabilities are planned for future updates.
Replacing GPT-3.5 Turbo in ChatGPT for Free, Plus, and Team users starting today.
It is currently accessible to developers and users of ChatGPT, and it is set to supplant GPT-3.5 for general applications.
The Significance: Its cost and capabilities make this low-end model much more within reach for most organizations looking to at least start off with some AI solutions.
TODAY’S SPONSOR
Use AI as Your Personal Assistant
Ready to embrace a new era of task delegation?
HubSpot’s highly anticipated AI Task Delegation Playbook is your key to supercharging your productivity and saving precious time.
Learn how to integrate AI technology into your processes, allowing you to optimize resource allocation and maximize output with precision and ease.
OpenAI Aspires to Develop Its Own AI Chip
The Veroneus: According to Reuters, OpenAI is in talks with chip designers like Broadcom to develop its own AI chip independent from the scare and costly GPUs of Nvidia.
The Details:
The OpenAI CEO, Sam Altman, has reached out to a semiconductor company called Broadcom to discuss the possible manufacturing of the chips in order to help it meet its data center requirements efficiently and increase its infrastructure capability.
They are also hiring former Google employees—AI hardware experts—to put their experience and expertise to work on the new chip, let alone establishing a factory network of such chips soon enough.
It was revealed that CEO Sam Altman is seeking billions of dollars to build a factory with Intel, aiming to produce semiconductors, and potential partners include Taiwan Semiconductor Manufacturing and Samsung Electronics.
The Significance: It is in pursuit of this that OpenAI is trying to ease the shortage of AI chips and boost the general AI industry, potentially benefiting everyone.
Mistral AI and Nvidia just introduced Mistral NeMo
The Veroneus: AI start-up Mistral has announced a partnership with NVIDIA to release an open-source, lightweight language model called Mistral NeMo. It is dubbed their best small model, but it will apparently be a big competitor against OpenAI's newly launched GPT-4o mini.
The Details:
Like GPT-4o mini, it can do more cost-efficient reasoning, world knowledge, and coding accuracy since NeMo requires less hardware compared to LLMs, hence cutting down infrastructure costs.
The performance of the model is strong in multi-turn conversations, mathematical, and common-sense reasoning, making it versatile for a wide range of enterprise applications.
NeMo is designed to support a single NVIDIA L40S, GeForce RTX 4090, or RTX 4500 GPU, which lets one run powerful AI on standard business hardware.
The Significance: NeMo is designed as a means of running LLMs at high performance but lower costs compared to larger LLMs, supporting multiple languages for global use, and democratizing powerful language models under an open-source umbrella.
Nvidia is Developing a New AI Chip for the Chinese Market
The Veroneus: According to Reuters, Nvidia is working on a new product codenamed B20, which is a more advanced version of the H20 chip for the Chinese market.
The Details:
Nvidia is allegedly working on a new chip that will comply with the USA's artificial intelligence chip export restrictions imposed on China.
According to allegations, the company partnered with Inspur, one of China's largest IT companies, for its chips codenamed B20.
Nvidia's most advanced chip for the Chinese market, the H20, had a weak start when deliveries began this year but managed to increase sales rates throughout the year.
The company, which offers chips to customers at lower prices than its rival Huawei in China, plans to sell more than 1 million H20 chips worth $12 billion this year.
The Significance: This development is crucial, as it strengthens NVIDIA's position in the Chinese AI market and is beneficial for market expansion, especially against competitors like Huawei.
🛑Hold on! Before you leave…
Your thoughts mean a lot! Please take a minute to give us your feedback.
What'd you think of today's edition? |
Your insights are invaluable to us. Please share any additional feedback you may have.