Google Gemma 3n

A dated reading of what is claimed, reported, and independently verified in the current evidence.

Current reading

Readiness band and full integer triple

AI-assisted assembly · derived results

Readiness band

Watch

Primary summary from verified readiness

Evidence strength Strong

Decision

Human editorial judgment · 2026-08-20

Adopt with guardrails

Why: Strong vendor backing, real multimodal on-device footprint (2-3GB), top-of-class sub-10B LMArena score, and a very large active community — but documented hallucination and reasoning limits mean it must be fenced, not trusted for open-ended reasoning.
Next: Pilot E2B/E4B via LiteRT-LM or Ollama on target devices; add factuality/grounding guardrails and constrain to bounded tasks (summaries, classification, voice ASR) where the reasoning ceiling and 30s audio limit are acceptable.

Constraints

No named blocker is present in the current public projection.

Evidence summary

AI-assisted assembly

Counts and dates only. Raw signals, private excerpts, trust records, and internal corpus material are not published here.

Publication record

Revision 12