Vertex AI Tuning

source

closed

training-code

closed

runs-on

Google-Cloud-Vertex-AI-only

license

Proprietary(managed GCP service)

Proprietary managed tuning service inside Google Cloud's Vertex AI generative-AI docs (now also mirrored under 'Gemini Enterprise Agent Platform' branding); no source, runs only on Google Cloud against closed Gemini models.

• https://cloud.google.com/terms/ recorded 2026-09-22

  Section 3.3 "Restrictions" of the Google Cloud Platform Terms of Service states Customer will not, and will not allow End Users to, "copy, modify, or create a derivative work of the Services", "reverse engineer, decompile, translate, disassemble, or otherwise attempt to extract any or all of the source code of, the Services", or "sell, resell, sublicense, transfer, or distribute any or all of the Services" — the standard proprietary-SaaS terms this closed, managed feature is offered under.

• https://docs.cloud.google.com/vertex-ai/generative-ai/docs/models/tune-models recorded 2026-09-22

  The tuning introduction page (redirected today to the canonical docs.cloud.google.com/gemini-enterprise-agent-platform/models/tuning) describes Gemini tuning entirely as a hosted capability: supervised fine-tuning, preference tuning, tuning checkpoints and continuous tuning are all run and served by Google, opening with "Agent Platform supports supervised fine-tuning to customize foundational models." No client library, SDK source, or self-hostable runtime is offered anywhere on the page.

Google publishes no usage figures for Gemini tuning, and its pricing page lists only per-token tuning rates. The Gemini app's monthly active users belong to a different product and cannot be attributed to this service, so the map leaves it unscored.

• https://cloud.google.com/vertex-ai/generative-ai/pricing recorded 2026-09-28

  The pricing page prices tuning per model: Gemini 3.5 Flash "Reinforcement Learning fine-tuning" is listed at "$0.01 / 1,000 count", and the per-1M-token table for "Gemini 2.0 Flash Lite" has a "Tuning for training tokens" row at "$0.001". A search of the body for monthly active, MAU, downloads and tuning jobs found none, so it publishes no adoption or usage figure for tuning.

• https://docs.cloud.google.com/vertex-ai/generative-ai/docs/models/tune-models recorded 2026-09-28

  The tuning guide is a how-to and reference page for Gemini tuning: it says "Agent Platform supports supervised fine-tuning to customize foundational models." Searched the body for customers, monthly active, downloads, MAU and tuning jobs: none, so no count of customers, developers, jobs or tokens tuned is published.

Vertex AI Tuning offers supervised, preference and preview reinforcement fine-tuning on Gemini models, with continuous tuning, checkpoints and function-calling tuning. Google exposes no controls over parallelism, precision or training scale and publishes no training benchmark, so it offers managed breadth rather than the deeper control of a training framework such as Megatron-LM.

• https://docs.cloud.google.com/vertex-ai/generative-ai/docs/models/tune-models recorded 2026-09-28

  Confirms Gemini tuning methods: "Gemini Enterprise Agent Platform preference tuning builds on supervised fine-tuning by letting you tune your Gemini models with human feedback data"; also a nav-listed "Reinforcement learning fine-tuning" section with a "Preview" badge, plus "Tuning checkpoints", "Continuous tuning" and "Tune function calling".