🏁 AI Race Timeline

OpenAI launches ChatGPT, starting the AI revolution. GPT-3.5 becomes the first widely available conversational AI.

GPT-4 sets a new standard for AI capability, passing the bar exam and medical licensing tests.

Anthropic releases Claude 2 with 100K context window, challenging GPT-4 in quality.

Google launches Gemini family, its most capable AI model, entering the frontier race.

Claude 3 Opus briefly tops leaderboards, proving Anthropic can match OpenAI at the frontier.

GPT-4o brings multimodal capabilities (text, vision, audio) to the leading model.

Meta releases the largest open-weight model, making frontier-class AI freely available.

OpenAI introduces reasoning models with o1, a new paradigm where AI "thinks" before answering.

Google launches Gemini 2.0 with improved speed and multimodal capabilities.

Chinese lab DeepSeek releases R1, matching frontier models at a fraction of the cost. Shakes the industry.

xAI's Grok 3 enters the frontier race, integrating with X (Twitter) platform.

Anthropic releases Claude Sonnet 4 with strong coding and reasoning performance.

Google Gemini 2.5 Pro brings thinking capabilities and tops several benchmarks.

OpenAI launches GPT-5, its most capable model, with significant reasoning improvements.

Anthropic Claude Opus 4.5 pushes the frontier in coding and agentic tasks.

DeepSeek V4 Pro matches frontier models while remaining significantly cheaper.

GPT-5.5 reaches Intelligence Index of 60, the highest measured score at launch.

Claude Opus 4.7 matches GPT-5.5 at the frontier with strong coding performance.

Google Gemini 3.1 Pro joins the frontier with top benchmark scores.

Moonshot AI Kimi K2.6 enters the top 10, showing China can compete at the frontier.

Anthropic Claude Opus 5 reaches the top of multiple leaderboards including LMArena.

OpenAI GPT-5.6 Sol joins Claude Opus 5 at the absolute frontier of AI capability.