Tech companies are constantly coming up with new AI models to stay competitive. The recent participant in the race is Mark Zuckerberg's Meta that has launched its latest generation of LLaMa (large language model) AI models -- Scout, Maverick and Behemoth -- as the competition in the AI race heats up between OpenAI, Google and other contenders.

According to Meta, these three models would enable users to build more “personalised multimodal experiences”. The company claimed that they outperformed Gemma 3, Gemini 2.0, Flash Lite and Mistral 3.1 across a wide range of benchmarks. Meta's blog post claimed that LLaMa 4 Scout is the world's best multimodal AI, capable of handling tasks from document summarisation to advanced reasoning across text, images, and video.

Further, Google is also enhancing its search-focused AI Mode chatbot by introducing multimodal capabilities, enabling it to interpret and respond to image-based queries. This update integrates a specialised version of Gemini AI with Google’s Lens image recognition technology, allowing users to upload or capture photos and receive detailed, informative responses, complete with relevant links. The multimodal features are available to a broader audience, accessible through the Google app on both Android and iOS devices.

Meanwhile, the recent launch of Meta's models followed closely on the heels of OpenAI's introduction of native image generation capabilities in ChatGPT. This new feature enables the chatbot to create more detailed, nuanced, and photorealistic images, impressing users across social media platforms. It sparked a trend of users generating Studio Ghibli-inspired images, showcasing the potential of AI to produce creative, high-quality visuals that capture the attention of diverse audiences.

Reasons for the ‘arms race’: This ‘arms race’ in AI innovation is happening for many reasons, such as the quest for dominance in emerging markets, responding to consumer demand, leveraging advancements in computing power, monetising data, future-proofing against disruptions, and addressing regulatory pressures. 

As Meta AI (LLaMa), OpenAI’s ChatGPT, and Google’s Gemini AI continue to compete with each other by introducing new models and features, let's compare how these AI systems perform. Here is the breakdown. 

ChatGPT (OpenAI): ChatGPT is one of the most well-known conversational AI models, with a focus on dialogue and human-like interaction. It powers everything from personal assistants to coding aids. ChatGPT has evolved through the GPT series, with the most recent being GPT-4, released in March 2023. GPT-4 offers better reasoning, creativity, and multi-turn conversation capabilities compared to earlier versions.

Focus: General-purpose conversational AI

Strengths:

• ChatGPT is designed to handle a wide range of queries and interactions with users in an engaging way.

• It's easy to interact with and accessible through various platforms (web, mobile, API).

• Fine-tuned for tasks like answering questions, creative writing, coding help, and brainstorming.

• OpenAI continues to improve its models (like GPT-4 and future versions), making them increasingly capable.

• ChatGPT is versatile across many domains—business, education, content creation, among others.

Weaknesses:

• While users can access it via Application Programming Interface (API) or chat interfaces, the model is not open for deep customisation without special access.

• Like any model, it can sometimes produce inaccurate responses.

Meta AI (LLaMA - Large Language Model Meta AI): Meta (formerly Facebook) has been investing heavily in AI, particularly focusing on large language models (LLMs), deep learning, and AI for social media and the metaverse. Meta’s LLaMA models have gained attention for their efficiency in both language generation and multimodal tasks.

Focus: Research-focused, open-access models

Strengths:

• Meta has made its LLaMA models available to researchers, providing flexibility for fine-tuning and adaptation in various applications.

• LLaMA is designed to be more flexible, allowing researchers to explore and experiment with model performance.

• LLaMA is part of Meta's larger efforts in AI research, focusing on advancing language modeling and ML techniques.

Weaknesses:

• LLaMA models aren’t specifically meant for conversational tasks. They may require more work to fine-tune for practical applications.

• It’s less user-friendly compared to ChatGPT in terms of daily interactions.

Google’s Gemini AI: Google’s Gemini is part of its advanced AI suite, evolving from the Bard AI chatbot. Gemini’s primary focus is on integrating Google’s extensive search capabilities with advanced AI for real-time info retrieval and large-scale multimodal understanding.

Focus: Google's new AI model family, aimed at integrating multimodal capabilities

Strengths:

• Gemini is designed to process not just text but also images and other inputs, making it more versatile than traditional text-only models.

• It benefits from Google’s vast data ecosystem, enhancing its ability to generate relevant and accurate answers based on a wide range of information.

• Google has a strong background in AI, and Gemini is expected to have advanced reasoning and knowledge capabilities.

Weaknesses:

• Gemini is still being developed and might not be as widely available or as refined for daily user interaction compared to ChatGPT.

• Google’s models are often integrated with their data services, raising questions about data privacy.

What to Choose: Each AI model has its strengths and weaknesses, depending on what you need—whether it’s polished conversational artificial intelligence, flexibility for research, or cutting-edge multimodal integration. The choice is yours. 

For AI innovation: OpenAI is currently the most advanced in terms of language and deep learning models (e.g., GPT series, DALL·E, etc.), with a focus on shaping the future of Genl AI.

For social media and VR: Meta has a significant lead in social media dominance (Facebook, Instagram, WhatsApp), and its investments in VR and AR with Oculus and the Metaverse are noteworthy. However, its future in these areas is uncertain.

For the overall tech ecosystem: Google is the most diversified and established with a robust presence in search, advertising, cloud computing, and hardware, according to experts. Google’s AI (DeepMind) is also pioneering, and it has vast resources to continue its growth.