Have you tried with frontier models from other providers, like Anthropic, Alphabet, xAI, etc? I am finding the capabilities with this specific test to be vary q

Have you tried with frontier models from other providers, like Anthropic, Alphabet, xAI, etc? I am finding the capabilities with this specific test to be vary quite wildly among "leading" models.