AI Model Landscape
Comparisons & Trends
Explore how AI models have scaled in size, expanded context windows, diversified architectures, and accelerated their release cadence — all visualized from our dataset of frontier models.
Parameter Count Over Time
Log-scale scatter plot showing how model sizes have evolved from millions to trillions of parameters. Dashed line shows the median trend.
Context Window Evolution
The explosion from 2K to 10M+ tokens — context windows have grown 5,000× in just a few years. Dashed line tracks the frontier maximum.
Architecture Distribution
How AI models are built — from decoder-only transformers to mixture-of-experts and diffusion models.
Modality Distribution
Text-only vs. multimodal vs. specialized — what types of data can these models process?
Release Cadence
Number of models released per quarter — showing the dramatic acceleration of AI model launches.
Models per Family
Number of tracked models in each AI model family, sorted by count.
Top Innovation Tags
Most common innovation tags across all models — the techniques and paradigms shaping modern AI.
Open vs. Closed Models by Year
Stacked bars showing the rise of open-weight and open-source models — the democratization of AI.
Research & Hardware
Papers that shaped the field and the hardware that made it possible
Research Paper Timeline
When landmark papers were published, colored by research era
Research Eras
Distribution of foundational papers across research eras
Innovation Adoption Over Time
How key innovations spread across models year by year
Hardware ↔ Model Co-evolution
How hardware accelerator releases enabled model breakthroughs