DeepSeek releases V4.1-Flash, says it outperforms flagship V4-Pro
Chinese artificial intelligence startup Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co. Ltd. today released DeepSeek-V4.1-Flash, the smallest model in a new architecture family. The company said tests by multiple parties put the open-weight model ahead of its much larger DeepSeek-V4-Pro on performance, cost, speed and total runtime. Starting Sept. 14, requests sent to…
Chinese AI startup DeepSeek has released V4.1-Flash, a new model that outperforms its flagship V4-Pro on multiple performance metrics. The open-weight model, with 552 billion parameters, is smaller than V4-Pro's 284 billion parameters and is billed at lower rates until V4.1-Pro is launched. The new model is a mixture-of-experts architecture with a causal encoder-decoder design, which keeps just 8 billion parameters active during processing.
DeepSeek claims that the model's memory efficiency, achieved through a four-bit floating-point format for key-value cache storage, results in a footprint of 890 bytes per token, about a quarter of what V4-Flash needs. The model scored 90.6 on Terminal-Bench 2.1, narrowly ahead of Anthropic's Claude Opus 5 at 89.1 and OpenAI's GPT-5.6 Sol at 88.8.
Brief written by urgent.news from SiliconANGLE's own syndicated text. Machine-written — may contain errors; check the original before relying on it.
