FriendliAI Highlights Performance-Focused Support for GLM-5.1 Model

According to a recent LinkedIn post from FriendliAI, the company is emphasizing its serverless endpoint support for GLM-5.1, an open model from Z.ai focused on software engineering tasks. The post cites third-party tracking sources Artificial Analysis and OpenRouter as indicating that FriendliAI ranks among the leaders in output speed, latency, and structured-output capabilities for this model.

The post also points to benchmark results suggesting GLM-5.1 can outperform Anthropic’s Claude Opus 4.6 on SWE-Bench Pro and CyberGym tests, while operating at a lower cost. For investors, this positioning could signal FriendliAI’s push to capture demand in high-value software engineering and agentic AI workloads, potentially enhancing its competitive standing in the inference and model-serving segment of the AI infrastructure market.