ernie-x1.1-preview

The Wenxin Large Model X1.1 demonstrates significant improvements in question answering, tool calling, agent capabilities, instruction following, logical reasoning, mathematics, and code tasks, with notable enhancements in factual accuracy.

The context length has been expanded to 64K tokens, enabling support for longer inputs and conversation histories while maintaining response speed and enhancing the coherence of long-chain reasoning.