The dates that explain how modern AI got here.

Grok 3 announced

Grok 3 marked xAI’s move into the reasoning-agent frontier with a much stronger model family.

DeepSeek-R1 released

DeepSeek-R1 pushed low-cost reasoning into the mainstream model conversation and accelerated global interest in Chinese frontier labs.

DeepSeek-V3 released

DeepSeek-V3 turned the lab into a serious global open-model competitor late in 2024.

OpenAI o1-preview launched

OpenAI’s o1-preview pushed reasoning-first models into the mainstream product conversation.

Claude 3.5 Sonnet launched

Claude 3.5 Sonnet quickly became one of the strongest coding and writing models in the market.

GPT-4o introduced

GPT-4o fused text, image and audio into a single flagship model and reset the baseline for mainstream multimodal products.

Llama 3 released

Llama 3 strengthened Meta’s position in open-weight models and raised the bar for broadly available open releases.

Claude 3 model family introduced

Claude 3 established Anthropic’s strongest model family to date and intensified the frontier race in 2024.

Gemma introduced

Gemma became Google’s main open-model family and widened the competitive open-weight field.

Gemini 1.5 introduced

Gemini 1.5 brought long-context performance back to the centre of the frontier model conversation.

xAI founded

xAI launched in 2023 and quickly became the lab behind the Grok model line.

Moonshot AI founded

Moonshot AI joined the post-ChatGPT Chinese lab wave and later built the Kimi family.

Estimated internal anchor: 1 Mar 2023. This is kept separate from the public canonical date.

DeepSeek founded

DeepSeek emerged in 2023 and later broke into the global frontier conversation with low-cost reasoning models.

Estimated internal anchor: 1 May 2023. This is kept separate from the public canonical date.

Mistral AI founded

Mistral AI launched in 2023 and quickly became Europe’s most prominent open-weight AI lab.

MiniMax founded

MiniMax was founded in early 2022 and became part of the new wave of Chinese multimodal model companies.

Estimated internal anchor: 1 Jan 2022. This is kept separate from the public canonical date.

Anthropic founded

Anthropic emerged as a safety-focused frontier lab and became the second major commercial model family of the ChatGPT era.

Cohere founded

Cohere launched with an enterprise-first strategy and became a durable commercial model provider.

Zhipu AI founded

Zhipu AI became one of the most visible Chinese frontier-model companies through the GLM line.

OpenAI founded

OpenAI launched as a non-profit AI research company and later became a central lab in the ChatGPT era.

DeepMind founded

DeepMind was founded in London and became one of the defining labs of the frontier-model era.