xAI Releases Grok 4.5; GLM-5.2 Benchmarks on SWE Bench Pro

Core summary

One sentence to understand this update

xAI released Grok 4.5, with GLM-5.2 appearing in its own benchmarks, showing 2.6 percentage points behind on SWE Bench Pro.

Impact & opportunity

What this could mean

The release of Grok 4.5 and the comparative SWE Bench Pro scores provide new data for evaluating coding-focused LLMs, influencing model selection for developers building code generation and analysis tools.