The introduction of the 3.5 Sonnet model marked a significant improvement in programming capabilities, enabling Cursor to enhance its features and user experience.

Sonnet 4 has effectively fixed previous deficits in coding models, showing significant improvements in intelligence and performance compared to earlier versions.

Anthropic's 3.5 Sonnet model was the critical breakthrough that made AI-generated code production-ready, unlocking the entire text-to-app market.

When we started using Claude Sonnet 3.5, we were impressed by its capabilities.

Opus is a significant advancement over Sonnet, as it understands my intent well and can one-shot many tasks that previous models struggled with.

By combining the intelligence of the 3.5 Sonnet model with custom retrieval models, Cursor can now perform multi-file edits, which has driven mass adoption of the platform.

Currently, Gemini excels in whole codebase indexing and planning, while Sonet 3.7 is the best for implementing code changes, although I found GPT 4.1 less effective recently.

The ability to influence the model's poetry creation before the final line is written provides strong evidence that it engages in advanced planning.

Recent evaluations of Claude 3.5 Sonnet included human evaluations where participants had to make business decisions based on extensive documentation, testing how an AI assistant could influence their choices.