Gemini / Google
The broadest full-stack AI platform on the market — frontier models plus real data infrastructure at Google scale; the tax is Google Cloud complexity and a product surface that reshuffles fast.
Element scores
Strengths
The only vendor covering effectively the whole periodic table with first-party products: frontier models (3.5 Flash, 3.1 Pro, Deep Think), a complete data layer (embeddings, Vector Search 2.0, RAG Engine, grounding with Google Search), the open-source ADK plus managed Agent Engine, the A2A protocol, real evaluation and observability services, Model Armor guardrails, and the widest multimodal generation surface (Veo 3.1, image, audio). Aggressive pricing at the small-model end and 900M+ app users give it unmatched distribution.
Honest dings
The product surface reshuffles constantly — Vertex AI was renamed Gemini Enterprise Agent Platform mid-2026, Agent Engine billing changed in January, and capabilities are split across AI Studio, the Cloud console and the Gemini app. Headline agentic features are gated: Gemini Spark is US-only in beta, Project Mariner and Genie sit behind the $199.99 Ultra tier. For small teams, the Google Cloud learning curve is real, and preview-to-GA churn makes today's integration tomorrow's migration.
Sources
Every audit lists the research it rests on — transparency and traceability are the product. Tools evolve: each audit is a snapshot of its audit date, and re-audits supersede older versions (kept below for reference).
- blog.google/innovation-and-ai/models-and-research/g… — Official (19/05/2026): Gemini 3.5 Flash GA, agentic benchmarks, Antigravity, Gemini Spark, Frontier Safety Framework and internal interpretability tooling (accessed 2026-07-07)
- blog.google/innovation-and-ai/models-and-research/g… — Official (19/02/2026): Gemini 3.1 Pro preview, ARC-AGI-2 77.1%, rollout across API/Vertex/app/NotebookLM (accessed 2026-07-07)
- blog.google/innovation-and-ai/models-and-research/g… — Official (03/03/2026): 3.1 Flash-Lite pricing $0.25/1M input, $1.50/1M output; thinking levels standard (accessed 2026-07-07)
- blog.google/innovation-and-ai/products/gemini-app/n… — Official (19/05/2026): 900M+ monthly users in 230 countries — the user count used in this passport (accessed 2026-07-07)
- blog.google/products-and-platforms/products/google-… — Official (19/05/2026): AI Ultra price cut $250→$200, new $99.99 5x tier, plan restructure (accessed 2026-07-07)
- gemini.google/ag/subscriptions — Official live pricing page: AI Plus $4.99, AI Pro $19.99, AI Ultra $99.99/$199.99; Veo 3.1, Deep Think, Spark tier gating (accessed 2026-07-07)
- docs.cloud.google.com/vertex-ai/docs — Official: Gemini Enterprise Agent Platform (formerly Vertex AI) — ADK, RAG Engine, Vector Search, Agent Gateway, agent evaluations, prompt optimizer, partner/open models (accessed 2026-07-07)
- docs.cloud.google.com/vertex-ai/docs/release-notes — Official release notes: Vector Search 2.0 GA (hybrid search, auto-embeddings), RAG Engine serverless, A2A on Agent Engine, tuning + eval integration, Memory Bank (accessed 2026-07-07)
- cloud.google.com/blog/products/ai-machine-learning/… — Official: A2A protocol (50+ partners), ADK open-source framework, MCP support, Agent Engine managed runtime, guardrails and tracing (accessed 2026-07-07)
- blog.google/company-news/inside-google/message-ceo/… — Official earnings (04/02/2026): Gemini app 750M MAU baseline, Gemini 3 momentum (accessed 2026-07-07)
- techcrunch.com/2026/02/04/googles-gemini-app-has-su… — Independent confirmation of the official 750M MAU disclosure (accessed 2026-07-07)
- techcabal.com/2026/06/05/google-ai-subscription-pla… — Independent (05/06/2026): plan restructure analysis, Ultra tier gating of Project Mariner/Genie, storage bundles (accessed 2026-07-07)