OpenRouter· @OpenRouter · X·· 3 小时前AI 评分45
AI 导读
它还能在单次响应中通过长解码写出最多 1M tokens，利用 continuation tokens 突破单次请求的输出上限。 详见我们的迁移指南：https://openrouter.ai/docs/cookbook/evaluate-and-optimize/model-migrations/gemini-4
正文
It can also write up to 1M tokens in a single response with long decode, using continuation tokens to run past the per-request output limit.
Learn more in our migration guide: https://openrouter.ai/docs/cookbook/evaluate-and-optimize/model-migrations/gemini-4
来源：OpenRouter · x.com