Another ultimate programming model is here: Claude Opus 4.5. But Anthropic also proved two things: - GPT-5.1 is slightly stronger in visual reasoning - Gemini 3 Pro excels in 'graduate-level reasoning difficulty' and 'multi-subject knowledge evaluation in multilingual environments' I get it now, these companies are finding their unique strengths, making it so you have to subscribe to multiple models to meet all your needs. Kind of like streaming subscriptions—each platform has its own must-watch content.