DeepSeek V4.1 Flash 1017发布，在Artificial Analysis评测中得分45分

DeepSeek于2026年10月17日发布了开源权重AI模型V4.1 Flash 1017（Max），该模型在Artificial Analysis Intelligence Index上得分45分，远超同类开源模型19分的中位数1。

该模型具有突出的性能指标。处理速度达到230.4 tokens/秒，首token时间延迟为1.08秒，上下文窗口容量为1M tokens1。模型由552B总参数组成，其中16B为活跃参数，支持文本和图像输入以及文本输出1。

在定价方面，输入费用为每百万tokens 0.30美元，输出费用为每百万tokens 1.20美元1。该模型采用MIT许可证发布，允许商业使用1，并可通过5个API提供商获取1。

DeepSeek released an open-source AI model called DeepSeek V4.1 Flash 1017 (Max) on October 17, 2026 1. The model achieved a score of 45 on the Artificial Analysis Intelligence Index, significantly outperforming the median score of 19 among comparable open-source models 1.

The model features a context window of 1.0 million tokens and processes information at a rate of 230.4 tokens per second, with a time-to-first-token latency of 1.08 seconds 1. It supports both text and image inputs while producing text outputs 1. The model comprises 552 billion total parameters with 16 billion active parameters 1. DeepSeek priced the model at $0.30 per million input tokens and $1.20 per million output tokens 1.

Released under an MIT license permitting commercial use, DeepSeek V4.1 Flash 1017 (Max) is available through five API providers 1.