Deep Dive

Models & Research

A chronicle of breakthrough model releases, groundbreaking research, and the relentless pursuit of artificial general intelligence.

14
Major Releases
4
Companies
3
Years Covered
Scroll to explore

The best way to predict the future is to invent it.

Alan Kay

2024

The year reasoning became real. o1 showed that AI could think step-by-step, while multimodal models merged text, image, and voice into unified intelligence.

November 20, 2024

Claude 3.5 Haiku Released

Anthropic releases Claude 3.5 Haiku, a faster and more cost-effective model that maintains strong performance while being significantly more affordable.

Anthropic Claude
🔥 May 13, 2024

GPT-4o Announced

OpenAI announces GPT-4o with native audio, vision, and text capabilities unified in a single model. The "o" stands for "omni."

OpenAI GPT-4 Multimodal
April 18, 2024

Llama 3 Released

Meta releases Llama 3 (8B and 70B) with improved performance and a new 8K context window, continuing the open-source revolution.

Meta Llama Open Source
🔥 March 4, 2024

Claude 3 Family Released

Anthropic releases Claude 3 (Haiku, Sonnet, Opus) with Opus achieving state-of-the-art on multiple benchmarks.

Anthropic Claude
February 15, 2024

Gemini 1.5 Pro: 1M Context

Google announces Gemini 1.5 Pro with a breakthrough 1 million token context window, enabling entirely new use cases.

Google Gemini Context

2023

The year of the foundation model wars. GPT-4 set a new bar, Llama 2 democratized open source, and multimodality went mainstream.

🔥 December 6, 2023

Google Gemini Launched

Google DeepMind launches Gemini, a natively multimodal model family trained from the ground up to understand text, images, audio, and video.

Google Gemini Multimodal
November 6, 2023

GPT-4 Turbo Announced

OpenAI announces GPT-4 Turbo with 128K context, lower pricing, and JSON mode at DevDay.

OpenAI GPT-4
🔥 July 18, 2023

Llama 2 Open-Sourced

Meta releases Llama 2 (7B, 13B, 70B) under a permissive license for research and commercial use, democratizing access to frontier-class models.

Meta Llama Open Source
July 11, 2023

Claude 2 Released

Anthropic releases Claude 2 with improved coding, math, and reasoning capabilities.

Anthropic Claude

2022

The year everything changed. ChatGPT arrived and the world would never look at AI the same way again.

Analysis

Key Insights

🏃

Accelerating Pace

Major model releases have increased from 1 in 2022 to 7 in 2024, with breakthrough capabilities arriving every few months.

🔀

Multimodal Convergence

By 2024, all major labs have shifted to multimodal-first architectures. Text-only models are becoming the exception.

🧠

Reasoning Revolution

o1's chain-of-thought approach marks a paradigm shift from pattern matching to deliberative reasoning.