PKU-ML/G1-7B
TEXT GENERATIONConcurrency Cost:1Model Size:7.6BQuant:FP8Ctx Length:32kLicense:apache-2.0Architecture:Transformer0.0K Open Weights Cold

PKU-ML/G1-7B is a 7.62 billion parameter causal language model, based on the Qwen2.5-Instruct architecture, specifically trained and fine-tuned using Group Relative Policy Optimization (GRPO) for advanced graph reasoning tasks. It demonstrates significant improvements on graph reasoning benchmarks like Erdős, while maintaining strong generalization to unseen graph tasks and preserving general reasoning abilities. This model is optimized for complex graph-related problem-solving and analysis.

Loading preview...