Google Gemini provides multimodal AI capabilities:
- Advanced reasoning with Gemini 2.5 and 3.0
- Multimodal understanding (text, image, video, audio)
- Video generation with Veo models
- Text-to-speech synthesis
- Cost-effective options with Flash variants
Available Models
Gemini 3 Pro
Next-generation model with enhanced capabilities and prompt caching.| Model | Input | Output | Cache Read | Best For |
|---|---|---|---|---|
| Gemini 3 Pro Preview | $2.60/M | $15.60/M | $0.26/M | Cutting-edge performance |
Gemini 3 Pro supports prompt caching - reuse context at 90% discount.
Gemini 2.5 Pro
Most capable current-gen models for complex tasks.| Model | Input | Output | Best For |
|---|---|---|---|
| Gemini 2.5 Pro | $1.62/M | $13.00/M | Production apps, complex reasoning |
| Gemini 2.5 Pro Preview (06-05) | $1.62/M | $13.00/M | Latest preview features |
| Gemini 2.5 Pro Preview (05-06) | $1.62/M | $13.00/M | Stable preview |
| Gemini 2.5 Pro Preview (03-25) | $1.62/M | $13.00/M | Earlier preview |
| Gemini 2.5 Pro Exp (03-25) | $1.62/M | $13.00/M | Experimental features |
Gemini 2.5 Flash
Fast, cost-efficient models for high-volume tasks.| Model | Input | Output | Best For |
|---|---|---|---|
| Gemini 2.5 Flash | $0.20/M | $3.25/M | Fast inference, production |
| Gemini 2.5 Flash Preview (05-20) | $0.20/M | $0.78/M | Latest features |
| Gemini 2.5 Flash Preview (04-17) | $0.20/M | $0.78/M | Standard preview |
| Gemini 2.5 Flash Preview (04-17 Thinking) | $0.20/M | $4.55/M | Extended reasoning mode |
| Gemini 2.5 Flash Lite Preview (06-17) | $0.13/M | $0.52/M | Ultra-lightweight |
Text-to-Speech
Gemini-powered voice synthesis models.| Model | Input | Output | Best For |
|---|---|---|---|
| Text-to-Speech 2.5 Pro | $1.30/M | $26.00/M | High-quality voice synthesis |
| Gemini 2.5 Pro Preview TTS | $1.30/M | $26.00/M | Preview TTS features |
| Text-to-Speech 2.5 Flash | $0.65/M | $13.00/M | Fast, cost-effective TTS |
| Gemini 2.5 Flash Preview TTS | $0.65/M | $13.00/M | Preview Flash TTS |
Pricing is per million characters of text input and audio output tokens.
Image Generation
Generate and understand images with Gemini.| Model | Input | Output (Text) | Output (Image) | Best For |
|---|---|---|---|---|
| Gemini 3 Pro Image | $2.00/M | $12.00/M | $120.00/M | High-quality image understanding & generation |
| Gemini 2.5 Flash Image | $0.30/M | - | $30.00/M | Fast, cost-effective image generation |
Image pricing is per million tokens.
- Image input:
560 tokens ($0.0011 per image with Gemini 3 Pro) - Image output 1K-2K:
1120 tokens ($0.134 per image with Gemini 3 Pro) - Image output 4K:
2000 tokens ($0.24 per image with Gemini 3 Pro) - Flash image output:
1290 tokens ($0.039 per image)
Video Generation (Veo)
Google’s text-to-video models for creating dynamic video content.| Model | Price | Per Minute | Best For |
|---|---|---|---|
| Veo 3.1 Generate Preview | $0.52/sec | $31.20/min | Latest video generation |
| Veo 3.1 Fast Generate Preview | $0.20/sec | $11.70/min | Fast video creation |
| Veo 3.0 Generate | $0.52/sec | $31.20/min | High-quality video |
| Veo 3.0 Fast Generate | $0.20/sec | $11.70/min | Balanced speed/quality |
| Veo 2.0 Generate | $0.46/sec | $27.30/min | Standard video generation |
Pricing is per second of video output. Example: A 30-second video with Veo 3.1 = 30 × $0.52 = $15.60
Best Practices
Model Selection
Use Gemini 3 for
- Cutting-edge features
- Latest capabilities
- Prompt caching needs
- Advanced reasoning
Use 2.5 Pro for
- Complex reasoning tasks
- Production applications
- High-quality outputs
- When accuracy matters
Use Flash for
- High-volume tasks
- Fast responses needed
- Cost-sensitive workloads
- Simple queries
Use Veo for
- Video generation
- Marketing content
- Creative projects
- Dynamic visuals
Prompt Caching (Gemini 3)
Optimize costs with prompt caching on Gemini 3 Pro:
- Cache Read: $0.26/M (90% cheaper than input)
- Use case: Repeated system prompts, documentation, knowledge bases
- Strategy: Place cacheable content at the start of your prompt
- 100K context without cache: $260/1M requests
- With cache: $26/1M requests = 90% savings
Context Windows
Gemini models support large context:- Gemini 3 Pro: Up to 2M tokens
- Gemini 2.5 Pro: Up to 2M tokens
- Gemini 2.5 Flash: Up to 1M tokens