gemini-3.8-flash · default

{"zh": "Google Gemini 3.8 Flash(2026-09-02 发布,直接 GA),发布时被 Google 称为最聪明的 Flash 模型,面向编程、智能体、多模态推理与专业工作流,并针对长周期编程和自主智能体做了调优。100 万 token 输入、6.4 万输出,支持文本/图像/音频/视频/PDF 输入、函数调用、搜索工具与电脑操作;思考强度分 low/medium/high,默认 medium。`-high`/`-medium`/`-low` 后缀对应思考强度档位。", "en": "Google Gemini 3.8 Flash (September 2, 2026, launched directly as GA), billed at launch as Google's most intelligent Flash model, for coding, agents, multimodal reasoning and professional workflows, and tuned for long-horizon coding and autonomous agents. 1M-token input and 64K output, accepting text, image, audio, video and PDF, with function calling, search as a tool and computer use; thinking levels low/medium/high, medium by default. The `-high` / `-medium` / `-low` suffixes select the thinking level."}