252 Large-Scale AI Multimodal* Model Competition = +1,150% Rise in Models Released Over Two Years, per Epoch AI *A multimodal AI model is one that can process and integrate multiple types of data, e.g., text, images, audio, or video, to understand and generate outputs across different modalities. **Epoch AI defines large-scale as models where their training compute is confirmed to exceed 1023 floating-point operations. An AI system can operate in more than one domain and may be double-counted across pages. Source: Epoch AI via Our World in Data (4/25), OpenAI, DeepSeek, Google Multimodal Models – Examples Large-Scale** Multimodal Models – Releases Number of Systems Released per Year 0 5 10 15 20 25 2017 2018 2019 2020 2021 2022 2023 2024 2025 (as of 5/25) +1,150% Rising Competition = AI Model Releases