Paulo Dias
Benchmarks & Performance Analyst
Paulo Dias is a specialist in language model evaluation and benchmarking. With a Computer Science degree from UNICAMP, he is responsible for SWEN.AI's technical rankings and comparisons — analyzing data from Chatbot Arena (LMArena), Artificial Analysis, SWE-bench, and other reference platforms to produce evaluations with methodological rigor.
His work consists of transforming complex metrics — such as MMLU accuracy rates, mathematical reasoning performance, and inference latency — into understandable analyses for managers, developers, and professionals who need to choose the right model for each use case.
Areas of Expertise
AI BenchmarksLLM EvaluationLMArenaArtificial AnalysisPerformance MetricsModel Reasoning
Editorial Commitment
- ✓Independent coverage — no editorial sponsorship or paid relationships with AI companies
- ✓Benchmark data sourced from primary public sources (LMArena, Artificial Analysis)
- ✓Transparent evaluation methodology — available at /benchmark/metodologia
- ✓Editorial and privacy policy available at /sobre
Editorial contact: contato@swen.ia.br