DeepSeek-V4.1-Flash-1017vsDeepSeek-V4.1-Flash
DeepSeek-V4.1-Flash-1017 与 DeepSeek-V4.1-Flash 有 9 个共同 benchmark，但两者公开成绩的测试模式不同，下方分数仅供参考，不判定胜负。
DeepSeek-V4.1-Flash-1017
DeepSeek-AI · 2026-10-17 · 多模态大模型
DeepSeek-V4.1-Flash
DeepSeek-AI · 2026-09-10 · 多模态大模型
测试模式不同的成绩
共 9 项。两款模型的公开成绩来自不同测试模式，分数不直接可比，不计入胜负和平均分差。
| 评测项 | DeepSeek-V4.1-Flash-1017 | DeepSeek-V4.1-Flash |
|---|---|---|
| Terminal-Bench 4.0 | 42.30Unspecified Reasoning (With Tools) | 26.80Max (With Tools) |
| Program Bench | 29.10Unspecified Reasoning (With Tools) | 20.30Max (With Tools) |
| Chartography | 83Unspecified Reasoning (With Tools) | 78.90Max (With Tools) |
| HLE | 42.60Reported best (effort unspecified) | 63.90Max (With Tools) |
| AutomationBench | 60.50Unspecified Reasoning (With Tools) | 54.80Max (With Tools) |
| DeepSWE | 77Unspecified Reasoning (With Tools) | 74.20Max (With Tools) |
| NL2Repo-Bench | 68.20Unspecified Reasoning (With Tools) | 65.40Max (With Tools) |
| BabyVision | 92.50Unspecified Reasoning (With Tools) | 89.60Max (With Tools) |
| ZeroBench Main | 51Unspecified Reasoning (With Tools) | 49Max (With Tools) |
规格对比
| 字段 | DeepSeek-V4.1-Flash-1017 | DeepSeek-V4.1-Flash |
|---|---|---|
| 发布机构 | DeepSeek-AI | DeepSeek-AI |
| 发布时间 | 2026-10-17 | 2026-09-10 |
| 模型类型 | 多模态大模型 | 多模态大模型 |
| 架构 | 稠密模型 | MoE 架构 |
| 参数规模 | 暂无数据 | 5520亿 |
| 上下文长度 | 1M | 1M |
| 最大输出 | 384K | 384K |
API 调用价格
价格优先使用 DataLearner 配置的 API 记录；缺失项不做推测。
| 价格项 | DeepSeek-V4.1-Flash-1017 | DeepSeek-V4.1-Flash |
|---|---|---|
| 文本输入 | $0.15 / 1M tokens | ¥1 / 1M tokens |
| 文本输出 | $0.6 / 1M tokens | ¥4 / 1M tokens |
| 缓存读取 | $0.003 / 1M tokens | ¥0.02 / 1M tokens |
小结
本页正文由结构化模型、价格与 benchmark 数据生成，不使用实时 LLM 撰写。