Last Updated on
ChatGPT, the worlds favorite AI chatbot, has a new rival. After just two months of training, xAI launched Grok AI on the 4th of November, 2023. However, GPT-4 and Grok-1 aren’t the only rivals in the ring, with xAI CEO Elon Musk and OpenAI CEO Sam Altman both ex-colleagues as board members of OpenAI in 2015. So, what’s the difference between ChatGPT vs Grok AI, and which is the better artificial intelligence?
ChatGPT vs Grok AI
Grok AI and ChatGPT have the most significant maturity gap between any two big tech AI chatbots. OpenAI’s ChatGPT was released to the public on November 30th, 2022. Now, almost one year later, Elon Musk responds with xAI’s Grok on November 4th, 2023. In this time, OpenAI CEO Sam Altman has spent an incalculable amount of $13 billion USD in funding from Microsoft to make his AI system the worlds favorite — a title it still retains.
However, Musk himself has a title his competitor doesn’t: World’s richest person. Having already spent $44 billion USD to acquire Twitter before rebranding it to X, Musk is now using one of the worlds largest and longest-running social media platforms as a training data source to catch up with OpenAI.
A unique and fundamental advantage of Grok is that it has real-time knowledge of the world via the 𝕏 platform. It will also answer spicy questions that are rejected by most other AI systems.
The xAI team
Grok AI is designed to be different. Elon Musk’s AI venture promises to bring humor and personality to what is an increasingly sanitized and censored form. To quote the xAI website, “Grok is designed to answer questions with a bit of wit and has a rebellious streak, so please don’t use it if you hate humor!”
Essential AI Tools
Winston AI detector
Originality AI detector
Jasper AI
WordAI
Copy.ai
AI Chatbot benchmarks comparison
xAI has conducted its own research into the proficiency of the Grok-1 LLM (large language model), compared with every other neural network in the AI chatbot space. A side-by-side comparison of Grok-1 puts it 4th highest across 4 separate independent benchmarks, despite having only two months of training. This puts it consistently above the free version of ChatGPT, which uses the GPT-3.5 model.
GSM8k is the first of these benchmarks. Based on “middle school math word problems”, this is a deceptively difficult task for a natural language processing (NLP) bot. This is because a neural network doesn’t use pure computational arithmetic like a calculator. Instead, it learns from example, concluding the answers to new problems based on successful input/output pairs (equation/solution in this case). Surprisingly, modern neural networks are getting very good despite this fundamental inefficiency, with GPT-4 scoring over 90%.
MMLU (Massive Multitask Language Understanding) is a multiple choice test. This would be easy for an AI, except for the fact that MMLU is multidisciplinary and tests the AI’s ability to relate from one field of expertise to another — exactly what neural networks should excel at.
HumanEval is a benchmark specialising in programming proficiency. It uses Python, the most popular programming language for creating these AI models in the first place, but extrapolates well when evaluating performance for other languages.
MATH is, unsurprisingly, a mathematics test. As shown by the performance figures below, these are much harder for an AI system than GSM8k, with the highest grade going to GPT-4 at merely 42.5%
| Benchmark | Grok-0 | LLaMa 2 | Inflection-1 | GPT-3.5 | Grok-1 | PaLM 2 | Claude-2 | GPT-4 |
|---|---|---|---|---|---|---|---|---|
| GSM8k | 56.8% | 56.8% | 62.9% | 57.1% | 62.9% | 80.7% | 88% | 92% |
| MMLU | 65.7% | 68.9% | 72.7% | 70.0% | 73.0% | 78% | 75% | 86.4% |
| HumanEval | 39.7% | 29.9% | 35.4% | 48.1% | 63.2% | N/A | 70% | 67% |
| MATH | 15.7% | 13.5% | 16.0% | 23.5% | 23.9% | 34.6% | N/A | 42.5% |
What’s the difference between ChatGPT and Grok AI?
Another key difference between Grok and ChatGPT is that, of course, there’s a waitlist to try Grok AI. You can sign up for the early access program now, but there is a manual approval process, whereas ChatGPT has been available without waitlist for several months now.
ChatGPT also has three different user tiers:
- Free
- ChatGPT Plus
- ChatGPT Enterprise
The free version of ChatGPT is free forever, meaning not merely a free trial. However, with a paid subscription to either of the two premium tiers, users gain access to the GPT-4 LLM, as well as other advanced features that Grok lacks, such as:
- GPT-4V (Computer vision-enabled VLM version of GPT-4)
- Browse with Bing (Internet access)
- Image generation (DALL·E 3)
- Plugins (943+ additional functions from third-party developers)
- Advanced Data Analysis (Previously Code Interpreter)
OpenAI vs xAI — AI chatbots compared
| Company | CEO | AI Chatbot | LLM | API | Open-source |
|---|---|---|---|---|---|
| xAI | Elon Musk | Grok | Grok-1 | No | No |
| OpenAI | Sam Altman | ChatGPT | GPT-3.5, GPT-4, GPT-4V, or GPT-4 Turbo | Yes | No |
| Sundar Pichai | Bard | PaLM 2 | Yes | No | |
| Microsoft | Satay Nadella | Bing Chat | GPT-4 | No | No |
| Meta | Mark Zuckerberg | Meta AI | LLaMA 2 | No | Yes |
| Anthropic | Dario Amodei | Claude | Claude-2 | Yes | No |
| Amazon | Andy Jassy | Olympus (rumored) | Olympus (rumored) | No | No |
Each of the companies above have publicly accessible AI chatbot technology. However, with the Tesla chief’s competitor still in early beta, overtaking the competition is nearly impossible task. Musks’ only saving grace appears to be his big data treasure trove that is X (formerly Twitter). Even in this, he’s not in a league of his own when compared to Google, which owns the worlds largest search engine, and Meta which runs not one but four of the worlds most used messaging and social media apps.