DeepSeek Releases v3.1 Model with Hybrid Reasoning Architecture

DeepSeek's V3.1 model introduces a hybrid architecture combining thinking and non-thinking modes. It features faster reasoning, improved tool use, and better multi-step task execution. Built on 839 billion tokens and using FP8 precision, the model supports a 128,000-token context length with 671 billion parameters. It ranks highly on benchmarks, scoring 71.6% on the Aider test—approaching GPT-4's performance at a fraction of the cost. Developers praise its cost/performance ratio and innovative hybrid inference approach.