Gemini 3 Pro Preview
API ID gemini-3-pro-preview
gemini-3-pro-preview was a preview Google model. It had a 1,048,576-token context window and up to 65,536 output tokens. It took image and PDF input and supported function calling, structured outputs and reasoning.
- Status
- Retired
- Released
- —
- Shutdown
- Mar 9, 2026 at the earliest
- Replacement
- Gemini 3.1 Pro Previewgemini-3.1-pro-preview
It was scheduled to shut down no earlier than March 9, 2026, which has passed. Check the provider's page before relying on it. Recommended replacement: gemini-3.1-pro-preview.
How to migrate to Gemini 3.1 Pro Preview
Benchmarks
Scores as each source published themComparing with Gemini 2.5 Pro, the previous model in this line. A score is compared only with the same benchmark from the same source, at the same setting where both have it. Full comparison: prices, limits, features
Arena
| Arena | Gemini 3 Pro Preview | Gemini 2.5 Pro | Change | Rank | Date | Source |
|---|---|---|---|---|---|---|
TextBy category
| 1485 ±4 | 1445 ±2 | +40 | 20 of 415 42,088 votes | Oct 13, 2026 | source |
| Text: Coding | 1517 ±7 | 1464 ±4 | +53 | 41 of 410 8,822 votes | Oct 13, 2026 | source |
| WebDev | 1440 ±8 | 1226 ±16 | +214 | 76 of 142 14,091 votes | Oct 13, 2026 | source |
| Vision | 1289 ±7 | 1248 ±5 | +41 | 13 of 165 13,639 votes | Oct 9, 2026 | source |
| Search | 1201 ±5 | 1144 ±5 | +57 | 18 of 42 38,857 votes | Oct 13, 2026 | source |
| Document | 1451 ±9 | 1431 ±6 | +20 | 31 of 44 10,769 votes | Sep 13, 2026 | source |
Rated as gemini-3-pro; Gemini 2.5 Pro as gemini-2.5-pro. Ratings come from people voting between two anonymous answers; Text, Vision, Search and Document are style-controlled, as on arena.ai. Data: Arena (arena.ai), CC BY 4.0.
Epoch AI
| Benchmark | Gemini 3 Pro Preview | Gemini 2.5 Pro | Change | Setting | Run | Source |
|---|---|---|---|---|---|---|
| GPQA Diamond | 92.6% | 85.3% | +7.3 pts | Nov 19, 2025 | source | |
| OTIS Mock AIME 2024-2025 | 91.4% | 84.7% | +6.7 pts | Nov 19, 2025 | source | |
| SWE-bench Verified | 72.9% | 57.6% | +15.3 pts | Feb 13, 2026 | source |
Run independently by Epoch AI. Data: Epoch AI, Capabilities & benchmarking (epoch.ai/benchmarks), CC BY.
Alias
All aliasesCurrently points to gemini-3.1-pro-preview (since March 9, 2026, per Google).
Features
Compare models| Feature | Value | Since | Source |
|---|---|---|---|
| Structured outputsThe model can be held to a JSON schema you supply, so every reply parses without retries or cleanup. | Supported | in place when tracking began (October 6, 2026) | source |
| Function callingAlso called tool use. You describe functions; the model replies with a function name and arguments for your code to run, then uses the result. | Supported | in place when tracking began (October 6, 2026) | source |
| Prompt cachingA prompt prefix you send again and again (system prompt, documents, tools) is cached, so later requests cost less and start faster. | Supported | in place when tracking began (October 6, 2026) | source |
| Batch APISend many requests as one job that runs in the background, usually finished within 24 hours, at a lower price than live requests. | Supported | in place when tracking began (October 6, 2026) | source |
| Image inputImages such as screenshots, photos and charts can go in the prompt for the model to read. | Supported | in place when tracking began (October 6, 2026) | source |
| PDF inputPDF files can go in the prompt directly, without converting them to text first. | Supported | in place when tracking began (October 6, 2026) | source |
| Reasoning / thinkingThe model can work through a problem step by step before it answers: usually better answers to hard questions, at the cost of more tokens and time. | Supported | in place when tracking began (October 6, 2026) | source |
| Built-in web searchThe company's API can search the web for the model during a request and cite its sources; you don't need a search service of your own. | Supported | in place when tracking began (October 6, 2026) | source |
| Context windowThe most the model can work with in one request, in tokens: your prompt, conversation history and files. A token is about three-quarters of an English word. | 1,048,576 tokens | in place when tracking began (October 6, 2026) | source |
| Max output tokensThe longest reply the model can write in one response, in tokens. A token is about three-quarters of an English word. | 65,536 tokens | in place when tracking began (October 6, 2026) | source |
Not published by Google: Fine-tuning, temperature parameter, top_p parameter, logprobs.
Price history
USD per 1M tokens, standard tierNo price history yet.
Prices
USD per 1M tokens, every type and band, by tierNo published prices.
Changelog
Changes we detected on Google's pagesNo changes recorded yet.
Last checked · Sources: Google deprecations, Google pricing, Google models, Google aliases, Google api models