Google offers several Gemini models, each optimized for different use cases. This guide helps you choose the right one for your Metorial-powered agent.

Available Models

Gemini 1.5 Pro

• Model ID: gemini-1.5-pro-latest
• Best for: Complex tasks, long context, function calling
• Context window: Up to 2 million tokens
• Recommended for: Most Metorial integrations

let response = await genAI.models.generateContent({
  model: 'gemini-1.5-pro-latest',
  contents: [/* ... */],
  config: { tools: session.tools }
});

Gemini 1.5 Flash

• Model ID: gemini-1.5-flash-latest
• Best for: Fast responses, high-volume applications
• Context window: Up to 1 million tokens
• Recommended for: Simple integration tasks, real-time applications

let response = await genAI.models.generateContent({
  model: 'gemini-1.5-flash-latest',
  contents: [/* ... */],
  config: { tools: session.tools }
});

Which Model Should You Use?

Use Gemini 1.5 Pro when:

• You need reliable function calling with complex integrations
• Your prompts require deep reasoning
• Context length is important (large documents, long conversations)
• Quality is more important than speed

Use Gemini 1.5 Flash when:

• You need fast responses
• You're handling high request volumes
• Your integration tasks are straightforward
• Cost efficiency is a priority

Testing Different Models

It's easy to test different models with Metorial:

const models = ['gemini-1.5-pro-latest', 'gemini-1.5-flash-latest'];

for (const model of models) {
  await metorial.withProviderSession(
    metorialGoogle,
    { serverDeployments: ['your-server-deployment-id'] },
    async session => {
      let response = await genAI.models.generateContent({
        model,
        contents: [{
          role: 'user',
          parts: [{ text: 'Test prompt' }]
        }],
        config: { tools: session.tools }
      });
      
      console.log(`Results from ${model}:`, response);
    }
  );
}

Recommendation

For most applications using Metorial integrations, start with Gemini 1.5 Pro. Its superior function calling capabilities work better with complex integration tasks. Switch to Flash only if you need faster responses and your integration tasks are simple.