Gemma 3n 2B is Google's smallest model in this batch, a free text-only model with an 8K context window and an intelligence index of just 1. Benchmark scores across the board are low: GPQA at 0.229, livecodebench at 0.095, and MMLU Pro at 0.378 confirm this is a model for simple, constrained tasks rather than business-grade reasoning.

For businesses, the practical use cases are narrow: lightweight on-device inference, simple classification, or prototyping where cost is the primary constraint. The 8K context window limits document handling, and there is no tool use or vision support.

As a free model from a trusted provider, it has value for experimentation and edge deployment. Teams should not use it for client-facing content, complex reasoning, or any workflow where answer quality matters.

Assessed June 30, 2026

Rankings consider pricing, capabilities, benchmarks, and real-world applicability and are refreshed as new models launch. Feedback?