“Salman Khan believes that leading LLMs from Google (Gemini), Anthropic (Claude), Groq, and OpenAI (GPT-4.5) are roughly equivalent in capability and can produce nearly identical behavior with proper prompting.”