#743·llm_wiki

模型在内部的思考(Chain-of-Thought)阶段耗尽了输出的 Token 限制,导致没有生成最终的正文内容

Author: PengGYingCreated Sep 10, 2026Updated Sep 14, 2026

Analysis failed: Model produced 14,735 characters of reasoning / chain-of-thought, but no actual response content. This usually means the endpoint hit a thinking-token limit, the model didn't transition from thinking to answering, or the endpoint is misbehaving (the official Anthropic / OpenAI APIs don't have this issue). Try a shorter input, increase max_tokens, or switch to a different model in Settings. 0001.md Analysis failed: Model produced 13,914 characters of reasoning / chain-of-thought, but no actual response content. This usually means the endpoint hit a thinking-token limit, the model didn't transition from thinking to answering, or the endpoint is misbehaving (the official Anthropic / OpenAI APIs don't have this issue). Try a shorter input, increase max_tokens, or switch to a different model in Settings.