DANTINOX: A Unified Framework for Multi-Paradigm Language Modeling

Language generation research increasingly spans three paradigms: autoregressive decoding, discrete masked diffusion, and continuous flow-matching. Comparing them is difficult because each typically lives in a separate codebase, meaning measured differences can reflect implementation details rather than the paradigms themselves.

DantinoX addresses this by providing an open-source JAX/Flax library in which a single modular Transformer backbone serves all three paradigms. Switching the generation paradigm, attention mechanism, or hardware topology requires only a configuration change, while the backbone architecture, tokenizer, initialization strategy, and training infrastructure remain consistent.

This design supports controlled cross-paradigm comparisons within one API for training, streaming inference, and benchmarking.