Home/Adoption Trends/AI Tool Launch Velocity: Releases Per Year
Adoption Trends

AI Tool Launch Velocity: Releases Per Year

The pace of AI model releases has accelerated dramatically. Epoch AI, the research organization that supplies Stanford's annual AI Index with its model tracking data, recorded 94 notable AI models released by industry in 2025 — more than ten times the count a decade earlier. Industry now produces over 90% of frontier models, up from under 50% in 2015. This page tracks model release velocity, the industry-versus-academia shift, the rise of open-weight models, and the changing geographic distribution of frontier AI research.

4 visualizations 5 sources Last updated June 2026 Free to embed
Loading chart...
CHART 1 · VELOCITY CURVE

Notable AI models released per year, 2015–2025

Loading chart...
Source: Stanford AI Index 2026 citing Epoch AI's notable model database. "Notable" defined per Epoch AI's published criteria: models that meet at least one threshold for highly cited research, state-of-the-art benchmark performance, or large training compute. Counts include both industry and other (academic, government) sources. The 2025 figure is 94 (87 industry + 7 other).
Chart 2 · Industry takes over

Industry vs academic share of notable models

Loading chart...
Source: Stanford AI Index 2024, 2025, 2026 reports citing Epoch AI. Industry share rose from under 50% in 2015 to over 90% in 2025. Academia's absolute output has remained roughly flat while industry's has grown dramatically.
CHART 3 · OPEN VS CLOSED

Open-weight share of new foundation models

Loading chart...
Source: Stanford AI Index 2024 and 2025 reports. Of 149 foundation models released in 2023, 65.7% were open-source. The trend reversed slightly in 2024 as frontier models moved closed, then accelerated again in 2025 with DeepSeek and other open-weight releases closing the performance gap with closed models (from 8% performance difference to 1.7% in one year per Stanford AI Index).
Chart 4 · Geographic distribution

Notable AI models by country of origin, 2024 vs 2025

Loading chart...
Source:  Stanford AI Index 2025 (US 40, China 15, Europe 3 in 2024) and Stanford AI Index 2026 (US ~50, China gaining, Europe ~5 in 2025). DeepSeek-R1 briefly matched the top US model in February 2025; as of March 2026 Anthropic's top model leads China's best by just 2.7% on benchmarks.

About this data

The AI Behavior Index is the research arm of OneChat AI. This page draws on the Stanford AI Index annual report, the most authoritative annual tracking of AI research and development trends, which itself cites Epoch AI's notable models database for the release counts.

One methodological caveat: the definition of "notable model" has been refined across editions of the AI Index. Stanford's 2024 report used "foundation models" (broader, 149 in 2023). The 2025 and 2026 reports shifted to Epoch AI's "notable models" (stricter, more comparable across years). Where this page shows counts before 2023, those are estimates based on retrospective Epoch AI database queries. The historical trend (acceleration) holds across both definitions. Data is refreshed annually when Stanford publishes the new AI Index edition. If you have a study to suggest or notice an error, contact us at research@aibehaviorindex.org.

For Journalists & Researchers

Use this data in your work.

Every statistic, chart, and graphic in this index is free to use and cite, with full source attribution. Can’t easily find what you need? Use our search bar to search by keyword, topic, or category.

✉️
Talk to our research team
Need a specific cut of data, an interview, or a quote? Email us — we typically respond within one business day.
research@aibehaviorindex.org →

How the data works

Every statistic shown is sourced from a publicly available study, survey, or report. We aggregate, organize, and contextualize this data — but the underlying research is conducted by the cited sources. Click any source link to access the original methodology. If you run into any issues or have a study to suggest, contact us at research@aibehaviorindex.org.