Welcome to Neural. AI is evolving rapidly, and we’re here to help you stay updated. OpenAI has announced that GPT-4.5 will soon be integrated into ChatGPT. But first, xAI is set to showcase Grok 3 today. The anticipation grows as we wait for OpenAI to make an unforgettable announcement…
Grok 3 Demo Reveal
Following Elon Musk’s bold suggestion to acquire OpenAI, he disclosed on Saturday that Grok 3 will be introduced tonight via a live demonstration at 8 p.m. PT.
Musk is generating buzz around Grok 3, labeling it the “smartest AI on the planet.” However, specific information about Grok 3 will remain under wraps until the announcement later today.
Here’s a brief timeline of xAI and Grok’s milestones:
- xAI launched in July 2023
- Grok-0 finalized in August 2023
- Grok-1 showcased in November 2023
- Grok-1.5 completed in March 2024
- Grok-2 released in August 2024
As of 2025, xAI has transformed Grok from just a feature of the X app to a fully independent application and website.
My foremost inquiry regarding Grok is: When will it be available in Tesla vehicles? While AI-enabled (supervised) self-driving functions are impressive, drivers could greatly benefit from a Grok-enhanced voice assistant that surpasses basic voice control capabilities.
The second question is whether OpenAI will overshadow xAI tonight with something exciting related to GPT-4.5.
OpenAI Updates This Week
Last week, Sam Altman from OpenAI outlined the upcoming roadmap for ChatGPT. Altman indicated that GPT-4.5 would debut in the coming weeks, making a release today seem improbable, while GPT-5 is anticipated shortly thereafter.
GPT-5 represents a more ambitious initiative, aiming to synthesize fast large language models with slower, reasoning-focused models, thus eliminating the need for multiple options per query.
In the meantime, Altman has been teasing the advantages of GPT-4.5 on X:
When asked whether he would “steal the spotlight” tonight, Altman responded diplomatically that “that wouldn’t be very nice…”
On a separate note, OpenAI has rolled out several real updates over recent days:
- OpenAI’s “o1 and o3-mini now support both file and image uploads in ChatGPT”
- OpenAI has increased o3-mini-high limits to 50 per day for Plus users — a 7x increase.
- OpenAI published a Model Spec update outlining desired behaviors for its AI models.
Additionally, there was this noteworthy vagueness in a recent hype post:
This update aims to enhance GPT-4o’s writing capabilities, particularly when provided with examples, as well as reducing the generation of low-quality output.
One thing is clear from recent model updates: GPT-5 will be vital for ChatGPT’s evolution. Understanding which models excel at various tasks has become increasingly complex and outdated before completion.
Anthropic’s Claude Update
In perhaps the most fascinating news of the past few days, Anthropic has announced plans to launch Claude 4 soon. Similar to GPT-5, Claude 4 is anticipated to be a unified chatbot offering rapid responses akin to Claude 3.5 while also managing more demanding reasoning tasks.
Claude 3 was launched in March, followed by Claude 3.5 in June, so the next significant update is expected. Keep an eye out for an official confirmation.
More News
In other AI developments, OpenAI’s board formally declined Elon Musk’s proposal to purchase the organization for $97.4 billion; Mira Murati’s startup has brought on yet another cofounder from OpenAI; France’s Mistral has launched a regional model focused on Arabic language and culture; AI search-oriented Perplexity has introduced Deep Research (unrelated to OpenAI’s similarly named offering); and Thomson Reuters successfully won the first major AI copyright case in the U.S.—ironically against the legal AI startup Ross Intelligence.
As we circle back, Grok 3 is trained on “all court cases,” which we will witness tonight.
Stay tuned for more updates on AI advancements in the next edition of Neural— exclusively on DMN! Access the previous issue here.
