deepseek-ai/DeepSeek-V3-0324

DeepSeek-V3-0324 is a 685 billion parameter language model developed by DeepSeek-AI, building upon the DeepSeek-V3 architecture. This iteration demonstrates significant improvements in reasoning capabilities across benchmarks like MMLU-Pro, GPQA, AIME, and LiveCodeBench. It is optimized for complex problem-solving, front-end web development, and enhanced Chinese writing proficiency, making it suitable for advanced analytical and creative tasks.

5.0 based on 3 reviews
Warm
Public
685B
FP8
32768
License: mit
Hugging Face