With each successive era of AI, from early scaling, to reasoning, to agentic, open models have taken half as long to catch up to the first closed model (SemiAnalysis)

Across successive LLM eras—early scaling, reasoning, and agentic—open-source models have required roughly half the time to reach performance parity with the leading closed-source model of each period. This trend indicates a accelerating convergence, narrowing the gap between open and proprietary AI systems.