GLM-5: New Leader in Open-Source Models

GLM-5 has established itself as the new leading open-source model on the Extended NYT Connections benchmark, achieving a score of 81.8. This result surpasses the previous high score of Kimi K2.5 Thinking, which had reached a score of 78.3.

The NYT Connections benchmark, available on GitHub, is used to evaluate the ability of language models to identify connections and relationships between concepts. GLM-5's performance suggests an improvement in the reasoning and natural language understanding capabilities of this model.

For those evaluating on-premise deployments, there are trade-offs to consider. AI-RADAR offers analytical frameworks on /llm-onpremise to evaluate these aspects.