Release archive
3.2.205

GEMINI-MODEL-USAGE-DOC-SYNC (AI Runtime) & GEMINI-PRODUCT-COPY-REFINEMENT (Dashboard + Marketing)

Release v3.2.205 focuses on GEMINI-MODEL-USAGE-DOC-SYNC (AI Runtime) and GEMINI-PRODUCT-COPY-REFINEMENT (Dashboard + Marketing).

Internal
GEMINI-MODEL-USAGE-DOC-SYNC (AI Runtime) & GEMINI-PRODUCT-COPY-REFINEMENT (Dashboard + Marketing)

Release digest

Internal

  • GEMINI-MODEL-USAGE-DOC-SYNC (AI Runtime) — Refreshed active repository, deployment, architecture, marketing-site, and public knowledge-base model references so current production usage is documented as the Gemini 3.x preview profile (gemini-3.5-flash fast paths and gemini-3.1-pro-preview report paths) with the Gemini 2.5 / 2.0 fallback chain preserved.
  • Updated active RAG wording to describe the Neon-backed retrieval runtime with Vertex embeddings instead of implying that managed Vertex RAG Engine is the current default retrieval path.
  • GEMINI-PRODUCT-COPY-REFINEMENT (Dashboard + Marketing) — Refined visible Gemini copy in the dashboard chat widget and public marketing components so user-facing UI names the Gemini intelligence layer without exposing internal model-profile wording.
  • Documentation decision: canonical project docs, internal runtime docs, marketing-site copy, public English docs, and localized knowledge-base copies were updated. Historical changelog/report/implementation artifacts were left unchanged because older Gemini 2.5 references there describe past runtime or translation events.

Technical notes