Gemini Diffusion vs InternS1

Gemini Diffusion

Google DeepMind

Gemini Diffusion is an experimental text and code generation model from Google DeepMind, announced at Google I/O in May 2025 as the first diffusion-based language model to achieve quality comparable to autoregressive models on standard benchmarks. Unlike transformer-based models that predict tokens sequentially left-to-right, it generates entire blocks of text by iteratively refining noise — the paradigm used in image and video generation models — enabling faster sampling speeds and stronger mid-generation error correction for code and mathematical editing tasks. At announcement it was available only as an experimental demo via waitlist, with no public API, marking it as a research milestone rather than a production deployment.

Shanghai AI Lab

InternS1, released by Shanghai AI Laboratory at WAIC 2025 on July 26, 2025, is a multimodal scientific reasoning large language model designed for advanced problem-solving across mathematics, physics, chemistry, and related domains. It supports text, image, and potentially other scientific data formats as input, and demonstrated strong performance on competition-level scientific benchmarks. InternS1 targets open-source scientific research, STEM education, and applications requiring deep domain reasoning across natural science disciplines.