Mapping the Efficiency Landscape of Small Language Models
This work evaluates 70+ SLMs from 2023–2025 on five task-specific benchmarks and compares them with two popular LLMs, revealing key trade-offs between energy, performance, and model selection and highlighting the need for informed, task-aware model selection rather than size-driven choices.
Fabian Reichwald, Lukas Schiesser, Christiane Plociennik et al.
· 0 citations