Gemini Pro

Gemini Pro refers to the maximum-capability tier within the Gemini family, a naming convention that started with the original Gemini 1.0 Pro in December 2023. At that time, Pro sat between Nano (on-device) and Ultra (maximum scale). Since Gemini 2 and 3, the lineup shifted to Flash / Flash-Lite / Pro, with Pro being the most capable generally available model. Gemini Ultra as a separate offering was retired and replaced by Deep Think mode within the Pro tier.

Gemini 3.1 Pro (launched February 2026) is the current generally available flagship as of May 2026. It runs at $2/$12 per million tokens for prompts under 200K context, with pricing doubling above 200K tokens. It supports a 1 million token context window, adaptive thinking (configurable reasoning depth), and integrated Google Search grounding. The model achieved a 2x jump over its predecessor on ARC-AGI-2 (77.1%) and leads on 13 of 16 Google-measured benchmarks. Gemini 3.5 Pro was announced at Google I/O 2026 with an expected rollout in June 2026.

For builders, the Gemini Pro tier is the choice when you need the strongest available reasoning from Google's stack: large codebases, complex document analysis, research synthesis over million-token contexts, or scientific reasoning tasks. The main practical consideration is the 200K token pricing cliff: for prompts frequently exceeding that threshold, context caching (up to 90% discount on cached tokens) becomes important for cost control.