API、托管或本地运行
DeepSeek-V4.1-Flash-1017
api、official_api
接入说明：DeepSeek API Models & Pricing (English) ↗选择要核对的信息，直接进入对应章节；各项事实均保留版本、条件和来源。
7/14 项有公开值 · 当前版本 DeepSeek-V4.1-Flash-1017 · 数据截止 2026-10-20
页面只展示已审核 Claim；待审核数据不会被当作公开事实。只展示已审核事实；未公开、冲突和过期状态不会被补猜。
只判断选型资料是否足够，不评价模型能力高低。
所有必填事实均通过状态、来源、证据与时效校验。最早核验时间：2026-10-20。
release_datelifecycle_statuscapability_tagsdeployment_optionscontext_window_tokensapi_idinput_price_usd_per_mtokoutput_price_usd_per_mtokavailabilityDeepSeek V4 Flash 是 DeepSeek 的 DeepSeek V4 模型，年鉴当前记录 2 个版本，最新公开记录为 DeepSeek-V4.1-Flash-1017。
权重为开放权重，代码为暂无已审核记录，训练数据为暂无已审核记录，部署方式为暂无已审核记录；这些维度分别判断，不能合并成单一“开源/闭源”结论。
当前公开快照的上下文窗口为 1M tokens，架构为 未公开；这里只使用通过审核且来源允许公开采集的 Claim。
6 个维度暂无可公开投影的审核记录
补充来源或线索会进入投稿队列，经编辑核验后才会成为公开事实。当前来源记录 3 个。
技术事实、业务价值、企业治理和证据质量分区呈现
厂商、家族、API ID、日期与生命周期
Dense/MoE、参数、层数、专家与 Tokenizer
预训练、SFT、偏好优化与安全对齐
教师/学生、基础模型、轨迹和限制
权重、代码、数据、许可证、部署与 API
输入输出、缓存、Batch 与部署成本边界
数据、权限、审计、合规、安全与人工兜底
19 条已审核事实 · 全部来自公开来源闸门通过的记录
尚无已审核公开记录的维度：02 能力与任务、07 部署与接入、09 性能与评测、11 业务场景、12 落地与提效。查看资料缺口 ↑
关系只来自年鉴数据库中已登记的归属字段。
下载、API、客户端和更新内容只展示已登记的来源事实；未登记不作推断。
api、official_api
接入说明：DeepSeek API Models & Pricing (English) ↗1 个已登记地区 · 1 条版本级记录
Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
保留发布日期精度、生命周期节点和原始来源，未公开日期不会被推断。
模型版本 · superseded
DeepSeek 官方 API 文档记录 v4-flash 已更新为 DeepSeek-V4-Flash-0731，调用方式保持不变；官网 2026-07-31 发布快照提供同日正式版 API 公测公告。
来源：DeepSeek API V4 Flash 0731 update ↗DeepSeek 官网原文称“DeepSeek-V4-Flash 正式版 API 已上线公测”，并强调 Agent 能力大幅增强；V4 Pro 暂未变动。
来源：DeepSeek official website ↗模型版本 · active
显示最新核验的 8 条 · 共 19 条已审核事实
Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
2 个版本 · 每项可回到来源与技术解释
查看最近版本差异高置信度，已绑定来源与证据。
条件：Cache hit
来源：DeepSeek API 模型与价格 ↗DeepSeek-V4-Flash-0731 and DeepSeek-V4-Pro: 1M context, maximum 384K output, thinking/non-thinking modes, JSON output, tool calls, API support, token prices, and concurrency limits.
高置信度，已绑定来源与证据。
来源：DeepSeek API 模型与价格 ↗DeepSeek-V4-Flash-0731 and DeepSeek-V4-Pro: 1M context, maximum 384K output, thinking/non-thinking modes, JSON output, tool calls, API support, token prices, and concurrency limits.
引用文本与接口返回都只包含已审核公开事实
DeepSeek DeepSeek V4 Flash｜最新版本 DeepSeek-V4.1-Flash-1017｜已审核事实 19 条｜公开来源 3 个｜数据截止 2026-10-20｜AI 年鉴｜https://ainianjian.com/models/model.deepseek.deepseek-v4-flash
公开接口返回该模型的版本、已审核事实 19 条及其来源信息，可直接用于脚本或对比。
https://api.ainianjian.com/api/models/model.deepseek.deepseek-v4-flash ↗curl -s 'https://api.ainianjian.com/api/models/model.deepseek.deepseek-v4-flash'
技术规格使用当前版本已审核且来源允许公开采集的事实；模型级开放信息与跨版本地区记录分别标注范围。请按每项适用范围核对，未登记字段显示「未公开」，不使用估算值填充。
Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
高置信度，已绑定来源与证据。
DeepSeek-V4-Flash-0731 and DeepSeek-V4-Pro: 1M context, maximum 384K output, thinking/non-thinking modes, JSON output, tool calls, API support, token prices, and concurrency limits.
高置信度，已绑定来源与证据。
条件：Cache miss
来源：DeepSeek API 模型与价格 ↗DeepSeek-V4-Flash-0731 and DeepSeek-V4-Pro: 1M context, maximum 384K output, thinking/non-thinking modes, JSON output, tool calls, API support, token prices, and concurrency limits.
高置信度，已绑定来源与证据。
来源：DeepSeek API 模型与价格 ↗DeepSeek-V4-Flash-0731 and DeepSeek-V4-Pro: 1M context, maximum 384K output, thinking/non-thinking modes, JSON output, tool calls, API support, token prices, and concurrency limits.
高置信度，已绑定来源与证据。
来源：DeepSeek API 模型与价格 ↗DeepSeek-V4-Flash-0731 and DeepSeek-V4-Pro: 1M context, maximum 384K output, thinking/non-thinking modes, JSON output, tool calls, API support, token prices, and concurrency limits.
高置信度，已绑定来源与证据。
来源：DeepSeek API 模型与价格 ↗DeepSeek-V4-Flash-0731 and DeepSeek-V4-Pro: 1M context, maximum 384K output, thinking/non-thinking modes, JSON output, tool calls, API support, token prices, and concurrency limits.
高置信度，已绑定来源与证据。
条件：The official page says the effective date is subject to a future announcement. Do not apply this multiplier to current prices.
来源：DeepSeek API 模型与价格 ↗DeepSeek states that the API will soon adopt a peak/off-peak pricing policy, with peak hours charged at 2x regular prices for all billing items; the effective date is subject to an official announcement. Peak hours are listed as 09:00-12:00 and 14:00-18:00 Beijing Time (UTC+8).
高置信度，已绑定来源与证据。
来源：DeepSeek API 模型与价格 ↗DeepSeek-V4-Flash-0731 and DeepSeek-V4-Pro: 1M context, maximum 384K output, thinking/non-thinking modes, JSON output, tool calls, API support, token prices, and concurrency limits.
高置信度，已绑定来源与证据。
来源：DeepSeek API 模型与价格 ↗DeepSeek-V4-Flash-0731 and DeepSeek-V4-Pro: 1M context, maximum 384K output, thinking/non-thinking modes, JSON output, tool calls, API support, token prices, and concurrency limits.
高置信度，已绑定来源与证据。
条件：Recorded on 2026-10-20; the pricing page states 'Use deepseek-flash as the model name' and that legacy names such as deepseek-v4-flash are mapped to this version.
来源：DeepSeek API Models & Pricing (English) ↗Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
高置信度，已绑定来源与证据。
条件：Set on 2026-10-20 from the pricing page feature table: vision is supported for deepseek-flash, tool calls and Responses API are supported, and the model supports thinking and non-thinking modes.
来源：DeepSeek API Models & Pricing (English) ↗Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
高置信度，已绑定来源与证据。
条件：Recorded on 2026-10-20 from the pricing page model details: context length 1M.
来源：DeepSeek API Models & Pricing (English) ↗Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
高置信度，已绑定来源与证据。
条件：Set on 2026-10-20 from the pricing page: the model is served through the DeepSeek API at https://api.deepseek.com in OpenAI and Anthropic compatible formats.
来源：DeepSeek API Models & Pricing (English) ↗Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
高置信度，已绑定来源与证据。
条件：Recorded on 2026-10-20 as the peak cache-miss input rate; off-peak is half of peak ($0.15) and peak hours are 01:00-04:00 and 06:00-10:00 UTC on weekdays excluding Chinese public holidays.
来源：DeepSeek API Models & Pricing (English) ↗Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
高置信度，已绑定来源与证据。
条件：Set on 2026-10-20 because the pricing page lists deepseek-flash as the current model name and the release note documents the 2026-10-17 version as the model now served by the API; the earlier DeepSeek-V4-Flash-0731 version is retired and its legacy alias is served by this version.
来源：DeepSeek API Models & Pricing (English) ↗Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
高置信度，已绑定来源与证据。
条件：Recorded on 2026-10-20 from the pricing page model details: max output 384K.
来源：DeepSeek API Models & Pricing (English) ↗Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
高置信度，已绑定来源与证据。
条件：Recorded on 2026-10-20 as the peak output rate; off-peak is half of peak ($0.60). The page lists cache-hit input separately at $0.006 peak and $0.003 off-peak per 1M tokens.
来源：DeepSeek API Models & Pricing (English) ↗Model details for deepseek-flash: model version DeepSeek-V4.1-Flash-1017; supports both non-thinking and thinking modes; context length 1M; max output 384K; features include JSON output, tool calls, Responses API, Anthropic API, chat prefix completion (beta), FIM completion (beta) and vision. Pricing per 1M tokens in USD: cache-hit input off-peak $0.003 and peak $0.006; cache-miss input off-peak $0.15 and peak $0.3; output off-peak $0.6 and peak $1.2; concurrency limit 2500. Notes state that deepseek-flash is the model name, off-peak rates are half of peak rates, peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday through Friday excluding Chinese public holidays, and the legacy name deepseek-v4-flash is still accepted but its model has been retired and is served by DeepSeek-V4.1-Flash-1017 at the Flash price.
高置信度，已绑定来源与证据。
条件：Set on 2026-10-20 from the official release note index entry 'DeepSeek-V4.1-Flash-1017 Release 2026/10/17'.
来源：DeepSeek-V4.1-Flash-1017 Release ↗The news index lists 'DeepSeek-V4.1-Flash-1017 Release 2026/10/17' and the release note introduces DeepSeek-V4.1-Flash-1017 as the current DeepSeek Flash model served through the DeepSeek API.