We put Claude 3.7 Sonnet's new "Extended Thinking" mode to the test with real-world math and coding challenges. The results? It solved a complex GMAT problem correctly while normal mode failed completely. Read our full breakdown on the XRay blog to see why this matters for your AI and automation strategy:
