miromind-ai/MiroThinker-32B-DPO-v0.1

MiroThinker-32B-DPO-v0.1 is a 32 billion parameter agentic language model developed by miromind-ai, built on Qwen3 architecture with a 32768 token context length. This DPO variant is designed for deep research and complex, long-horizon problem-solving, integrating advanced capabilities like task decomposition, multi-hop reasoning, and retrieval-augmented generation. It achieves state-of-the-art performance among open-source models on the GAIA benchmark, demonstrating strong capabilities in real-world agentic tasks.