Models
Information about Google Gemini models supported in Woozlit.
Google DeepMind develops the Gemini series, a multimodal-native family of models.
Owner
Google (Alphabet Inc.), led by Sundar Pichai.
Models Available (2025)
- Gemini 2.5 / 3.0 Series: (Roadmap) Next-generation models focusing on extreme context and reasoning.
- Gemini 2.0 Flash: The new default standard for speed and multimodal performance.
- Gemini 1.5 Pro: Features a massive 2 Million token context window.
- Gemini 1.5 Flash: Highly efficient, low-latency model for high-volume tasks.
Pricing
| Model | Input / 1M Tokens | Output / 1M Tokens |
|---|---|---|
| Gemini 1.5 Pro | $3.50 | $10.50 |
| Gemini 1.5 Flash | $0.075 | $0.30 |
| Gemini 2.0 Flash | Preview pricing | Preview pricing |
(Pricing scales with context length; >128k context is higher set)
Performance
Gemini models excel at long-context understanding (documents, codebases, videos).
- 2.0 Flash: Multimodal native (audio/video/text) with incredibly low latency.
- 1.5 Pro: Best for "needle in a haystack" retrieval tasks across thousands of pages.
Privacy
Does Google train on my API data? No. When using the paid Gemini API (Vertex AI or paid AI Studio quota), Google does not use your data to train its models.
- Free Tier Warning: Data submitted to the free tier of AI Studio or Gemini Apps may be used for training and human review. Always use the paid API for enterprise privacy.