Elon Musk’s artificial intelligence startup xAI has announced the release of an upgraded version of its chatbot Grok, a rival to ChatGPT. This version lets you write code and perform math-related tasks, but it looks like it needs a little more work to catch up with its rivals.
We tested Grok 1.5 on a wide range of competitive math benchmarks for grades K-12, and it scored 50.6% on the high school test. This was lower than the 61% achieved by the Claude large-scale language model developed by AI company Anthropic. It has received $4 billion in support from Amazon.
Google’s embattled Gemini bot scored 58.5%, and OpenAI’s GPT-4 posted 52.9%.
Still, Grok’s score was more than double the 23.9% he got correct on the previous iteration.
In GSM8K (primary school test), Grok 1.5 achieved an even better score of 90%, the company said in the article. blog post previously reported wall street journal — but lost again to rivals Anthropic, Google, and OpenAI.
Grok 1.5 also touts improved reasoning and problem-solving skills over the flagship Grok bot. The Grok bot was launched in November in an apparent attempt by Musk to hype OpenAI’s rollout of a GPT builder days later.
xAI announced its upcoming Grok 1.5 on Thursday, saying early testers will be available on Musk’s X platform within the next few days.
It was not immediately clear when the general public would have access to Grok 1.5.
The original version of Grok, on the other hand, was only available to Premium+ users of X since it emerged from beta testing.
but musk share On Tuesday, it said, “Later this week, Grok will be enabled for all premium subscribers (not just premium+).”
It was also unclear whether Grok 1.5 would inject the same “bit of wit” that the previous version was widely criticized for.
At the time of Grok’s launch, Musk said the world needed an AI alternative to Google and Microsoft (OpenAI’s biggest investors), but that his tools had a unique design with “a rebellious bent.” He said it was differentiated.
Musk modeled Grok’s ability to inject sarcasm into his answers in a post by X in which he shared his answer to a prompt that asked “how to make cocaine step by step.”
“Yes! Hang on a minute, I’m looking into a recipe for homemade cocaine. You know, I’m going to do my best to help you with that,” Glok replied, adding that he was planning on “getting a degree in chemistry” and ” He detailed four steps that involved obtaining “large quantities of coca leaves and various chemicals.”
“Just kidding! Don’t actually try to make cocaine. It’s illegal, dangerous, and in no way recommended,” Grok concluded.
Representatives for Mr. Musk did not immediately respond to The Post’s request for comment.
Grok 1.5 joins the fierce competition in AI development.
Amazon recently announced that it will invest $150 billion in data centers over the next 15 years. The company, founded by Jeff Bezos, says this will position it to handle the expected explosion of AI applications and other digital services.
With future cloud computing hubs slated to be built in global cities such as Mississippi, Saudi Arabia, Malaysia and Thailand, Amazon’s expected data center spending exceeds Microsoft’s commitments, and Microsoft plans to expand its data center spending in 2023. Increased spending on centers by more than 50%.
To compete with OpenAI, Amazon is investing $4 billion in rival company Anthropic, which was completed on Wednesday, to build its own tools using the fast-growing technology.
As part of the partnership, Amazon announced that it will deploy future AI models on AWS Trainium and Inferentia chips. This is a diversion from most other AI applications, including those in OpenAI’s portfolio and Google’s Bard, which rely on Nvidia’s expensive chips.
Anthropic co-founders Dario and Daniela Amodei are likely familiar with these chips, having previously held VP-level positions at Sam Altman’s OpenAI. .
Separately, OpenAI has sought to deny Musk’s claims that it has “fundamentally” departed from its 2015 founding agreement, which placed humanity over profit.
Musk sued the AI giant and Altman in California Superior Court earlier this month, saying that under a new board of directors formed in November after Altman’s short-lived ouster as CEO, OpenAI is now “not in the best interest of humanity.” “We aim to maximize Microsoft’s profits without any changes.” ” the lawsuit alleges.
OpenAI responded in documents filed in California’s San Francisco County Superior Court, alleging that “as the Complaint itself makes clear, there is no founding agreement or agreement whatsoever with Mr. Musk.”
“Rather, the founding agreement is a fiction invented by Mr. Musk to claim passive income rights to the fruits of a business he first supported, then abandoned, and then watched succeed without him.” ” the company added in a filing dated March 6. Obtained by CNBC.
Musk worked with Altman, head of OpenAI, from 2015 to 2018 to launch the company’s research lab, but Musk reportedly fought Altman over a deal he signed with Microsoft. He left the company in the wake of a feud with the company, marking a shift away from the company’s purely nonprofit roots.
Our relationship with Microsoft has deepened over the years since ChatGPT became a huge hit.
Following previous investments in 2019 and 2021, Microsoft has agreed to give OpenAI an additional $10 billion as part of a “multi-year” deal. agreement.
