The Latest Advances in AI Models

The rapid development and release of AI models by tech giants like Google and startups such as OpenAI and Anthropic can be overwhelming to keep track of. The promotion of these models is often based on industry benchmarks, which may not accurately reflect their real-world applications and usability.

Cutting Through the Noise

To provide clarity, TechCrunch has compiled a comprehensive overview of the most advanced AI models released since 2024. This list includes details on how to utilize these models and their ideal applications. The list will be regularly updated to reflect the latest launches.

The Sheer Volume of AI Models

With over 900,000 AI models hosted on platforms like HuggingFace, it’s possible that some models may perform better in specific ways. However, this list aims to provide a concise and informative overview of the most notable AI models.

AI Models Released in 2025

• OpenAI o3-mini: A reasoning model optimized for STEM-related tasks, available for free with optional subscription for heavy users.
• OpenAI Deep Research: A service designed for in-depth research with clear citations, available with ChatGPT’s $200 per month Pro subscription.
• Mistral Le Chat: A multimodal AI personal assistant with a paid version featuring up-to-date journalism from the AFP.
• OpenAI Operator: A personal intern AI that can perform tasks independently, available with ChatGPT’s $200 per month Pro subscription.
• Google Gemini 2.0 Pro Experimental: A flagship model that excels at coding and general knowledge, requiring a Google One AI Premium subscription of $19.99 a month.

AI Models Released in 2024

• DeepSeek R1: A Chinese AI model that performs well on coding and math, available for free with open-source capabilities.
• Gemini Deep Research: A service that summarizes Google’s search results in a simple and well-cited document, available with a $19.99 Google One AI Premium subscription.
• Meta Llama 3.3 7B: The newest and most advanced version of Meta’s open-source Llama AI models, available for free.
• OpenAI Sora: A model that creates realistic videos based on text, available on paid versions of ChatGPT starting at $20 a month.
• Alibaba Qwen QwQ-32B-Preview: A model that excels in math and coding, available for free with open-source capabilities.
• Anthropic’s Computer Use: A service that allows AI to control your computer to complete tasks, available in beta with pricing via API.
• x.AI’s Grok 2: An enhanced version of x.AI’s flagship chatbot, available with free and subscription-based options.
• OpenAI o1: A model that produces better answers through a hidden reasoning feature, available with ChatGPT Plus subscription.
• Anthropic’s Claude Sonnet 3.5: A model known for its coding capabilities, available for free with optional Pro subscription.
• OpenAI GPT 4o-mini: A model touted as the most affordable and fastest yet, available on ChatGPT’s free tier.
• Cohere Command R+: A model that excels at complex Retrieval-Augmented Generation applications for enterprises, available for enterprise users.