Science ❯ Computer Science ❯ Artificial Intelligence ❯ Model Evaluation

Performance Metrics

MoE Research Delivers New Playbooks for Design, Routing, and Real-World Deployment

The releases point to simpler design rules with practical fixes that make sparse expert models easier to scale.

OpenAI Rolls Out GPT-5.5, Shifting ChatGPT From Assistant to Digital Worker

LLM-Agent Research Coalesces Around Generalizability, RLVR, and Grounded Retrieval

September Research Blitz Reveals LLM Vulnerabilities and Rolls Out Real-World Benchmarks

OpenAI Launches o3-Pro, Its Most Advanced AI Reasoning Model Yet