April brought several notable foundation model releases with longer context windows and stronger reasoning benchmarks. Open-weight alternatives continue to narrow the gap with closed models.
Developers are testing the new models on retrieval-heavy and agentic tasks. Vendors emphasize safety evaluations and responsible use guidance.
