Model
60d agoClaude Sonnet 5 Outperforms Opus in Knowledge Work
A score of 1,618 on the GDPval-AA v2 test puts Claude Sonnet 5 ahead of the larger Opus 4.8. Anthropic designed the model to beat its predecessor across all benchmarks. It remains intentionally limited in cybersecurity tasks to avoid government blocks.
The Signal
This narrows the performance gap between mid-tier and premium models.