Bursa malaysia: H2O.ai, a pioneer in sovereign AI and a leader in highly accurate predictive AI, today announced that its openly available H2OVL Mississippi visual-language multimodal models have surpassed one million downloads monthly. This significant milestone coincides with the global AI community's anticipation of NVIDIA GTC 2026.

According to BERNAMA News Agency, the H2OVL Mississippi models, hosted on Hugging Face, have quickly become one of the most adopted open multimodal OCR-capable models in the ecosystem. They have matched leading category performers such as Qwen, DeepSeek, and Google's Gemma, as publicly benchmarked. This increase in adoption underscores a growing trend towards sovereign, open, transparent, and locally deployable AI models-a direction that H2O.ai has been advocating for over a decade.