DeepSeek V4 Flash
DeepSeek released the 文本 model · Launched on 2026-07
Open Source
文本
开源
API 可用
模型简介
DeepSeek V4 Flash is the latest open-source model released by DeepSeek on July 31, 2026, with a 284B total parameter MoE architecture, activating only 13B parameters for inference. It has a 1 million token context, and API input cost is as low as ¥0.02/million tokens after cache hit. Performance is comparable to GPT-5, Agent capabilities surpass Claude Opus 4.8, and Unsloth released a local running version on the same day. This model was retired on September 10, 2026 with the release of DeepSeek-V4.1-Flash; existing API requests are now served by V4.1-Flash and billed at Flash rates.
核心特性
- 284B MoE 架构
- 100万 tokens 上下文
- 385K 最大输出
- ¥0.02/M 极低 API 成本
核心优势
- API 成本最低
- 性能对标 GPT-5
- Agent 能力突出
- 开源可商用
技术规格
参数量
284B (13B 激活)
上下文窗口
1M tokens
API 支持
Yes
开源状态
Open Source
发布日期
2026-07
提供商
DeepSeek
价格信息
输入 ¥0.14/1M tokens（缓存命中 ¥0.02），输出 ¥0.56/1M tokens；开源免费
基准测试
性能对标 GPT-5, Agent 基准超越 Opus 4.8, 284B MoE 架构业界领先