A large open-weight base model from ByteDance-Seed with a notably expansive 524K token context window, meaning it can hold and process very long documents or conversations in a single pass. As a base model, it hasn't been fine-tuned for chat or instruction-following, so it behaves more like a raw text predictor than a conversational assistant — it's a foundation meant for further training or adaptation. Its Apache 2.0 license makes it freely usable and modifiable.