TrajectoryCapabilityThe race for the most capable AI

UpdatedyesterdayBeats8

The race for the most capable AI

For two and a half years, the most-watched question in AI has been which company has built the most capable model. The answer has shifted multiple times — from OpenAI's runaway lead in 2023, to a real two-horse race with Anthropic in 2024–25, to a four-way contest with Google and xAI in the mix today. Whoever holds the lead in any given quarter sets the prices, the standards, and the direction every other AI company has to react to.

Timeline

  1. March 14, 2023

    OpenAI launches GPT-4. The gap to every other AI company is wide enough that for most of the year, GPT-4 has no real competitor.

  2. September 25, 2023

    Amazon commits up to $4 billion to Anthropic. The deal signals that the race is going to be a contest between well-capitalised labs, not a runaway by a single company.

  3. June 21, 2024

    Anthropic ships Claude 3.5 Sonnet. For the first time, a non-OpenAI model is widely regarded as better at coding — the first dimension on which the lead changed hands.

  4. September 12, 2024

    OpenAI ships o1, a model that pauses to "reason" before answering. The frontier opens a new dimension of capability, beyond chat-style responses.

  5. February 24, 2025

    Anthropic ships Claude 3.7 Sonnet with reasoning built in. Anthropic's coding lead deepens; the ranking depends on which task you measure.

  6. December 11, 2025

    Google's Gemini 2.5 reaches parity with OpenAI and Anthropic on most benchmarks. The race becomes genuinely four-way — OpenAI, Anthropic, Google, xAI.

  7. April 22, 2026

    Anthropic overtakes OpenAI in business-software deployment, per Ramp's enterprise spending data — the first time enterprise customers are spending more on Anthropic than on OpenAI.

    Source: Ramp AI Index

  8. May 15, 2026

    OpenAI launches GPT-5. The new model retakes the top spot on multiple benchmarks, but the lead is measured in weeks, not the years it once was.

Where things stand right now

After a year of Anthropic catching up and overtaking OpenAI on several dimensions, GPT-5's May launch has put OpenAI back on top — but barely, and only on some benchmarks. Anthropic still leads in business deployment, Google still leads in cost-per-query, and the era of any single company being clearly ahead is over.