K2-V2: A 360-Open, Reasoning-Enhanced LLM
This provides a reasoning-enhanced, fully open LLM for the AI community, though it is incremental as it builds on existing LLM paradigms with specific enhancements.
The authors tackled the problem of creating a superior open-source large language model (LLM) for reasoning adaptation, resulting in K2-V2, which rivals open-weight leaders in its size class, outperforms Qwen2.5-72B and approaches Qwen3-235B in performance.
We introduce K2-V2, a 360-open LLM built from scratch as a superior base for reasoning adaptation, in addition to functions such as conversation and knowledge retrieval from general LLMs. It stands as the strongest fully open model, rivals open-weight leaders in its size class, outperforms Qwen2.5-72B and approaches the performance of Qwen3-235B. We actively infuse domain knowledge, reasoning, long-context, and tool use throughout the training process. This explicitly prepares the model for complex reasoning tasks. We demonstrate this potential using simple supervised fine-tuning, establishing a strong baseline that indicates significant headroom for advanced alignment. By releasing the full training history and data composition, we maximize the effectiveness of continuous training, a key open source production scenario. We release the model weights and signature LLM360 artifacts, such as complete training data, to empower the community with a capable, reasoning-centric foundation.