Skip to main content

Explore Google's AI: Gemini & Gemma Series Overview

Google Gemini
info

For detailed information about using models with APIpie, check out our Models Overview and Completions Guide.

Description​

Google's AI models, including the Gemini Series and Gemma Series, represent cutting-edge advancements in artificial intelligence. Developed by Google DeepMind, these models leverage state-of-the-art technology to deliver exceptional performance across various AI tasks. The models are available through multiple providers integrated with APIpie's routing system.

The Gemini Series represents Google's most advanced AI models, capable of sophisticated reasoning across text, code, images, and video. Meanwhile, the Gemma Series offers efficient, open-source models built on the same research and technology as Gemini. These models are accessible through Google Cloud's Vertex AI platform and various third-party providers.

Key Features​

  • Advanced Token Processing: Models support context lengths from 4K to 2M tokens for comprehensive text processing, with optimized performance across different scales.
  • Multi-Provider Access: Available through platforms like Google Cloud, OpenRouter, EdenAI, and more, with enterprise-grade security.
  • Versatile Applications: Optimized for chat, instruction-following, code generation, and multimodal tasks, with state-of-the-art performance across domains.

Model List in the Google Series​

Model List updates dynamically please see the Models Route for the up to date list of models​

info

For information about model performance and benchmarks, see the Gemini Technical Report and Gemma Technical Report.

Model NameMax TokensResponse TokensProvidersSubtype
gemini-pro-1.52,000,0008,192OpenRouterText
gemini-flash-1.51,000,0008,192OpenRouterChat
gemini-flash-1.5-8b1,000,0008,192OpenRouterChat
gemini-pro91,72822,937EdenAIChat
palm-2-codechat-bison-32k91,75022,937OpenRouterCode
palm-2-chat-bison-32k32,7688,192OpenRouterChat
gemini-pro32,7608,192OpenRouterText
palm-2-codechat-bison20,0702,867OpenRouterCode
palm-2-chat-bison9,2161,024OpenRouterChat
gemma-7b-it8,1928,192DeepinfraChat
gemma-2-27b-it8,1928,192TogetherChat
gemma-2b-it8,1928,192TogetherChat
gemma-2-27b-it8,1924,096OpenRouterText
gemma-2-9b-it4,0964,096OpenRouter, Monster, TogetherChat
gemini-pro-vision16,3842,048OpenRouter, EdenAIVision
gemini-1.5-flash--EdenAIChat
gemini-1.5-pro--EdenAIChat
gemini-1.5-flash-latest--EdenAIChat
gemini-1.5-pro-exp-0801--EdenAIChat
gemini-1.5-pro-exp-0827--EdenAIChat
gemini-1.5-pro-latest--EdenAIChat
gemini-1.5-flash-8b--EdenAIChat
chat-bison--EdenAIChat
gemma-1.1-7b-it--DeepinfraChat
vit-base-patch16-224--DeepinfraImage
vit-base-patch16-384--DeepinfraImage
textembedding-gecko768-EdenAIEmbedding

Example API Call​

Below is an example of how to use the Chat Completions API to interact with a model from the Google Series, such as gemini-pro-1.5.

curl -L -X POST 'https://apipie.ai/v1/chat/completions' \
-H 'Content-Type: application/json' \
-H 'Accept: application/json' \
-H 'Authorization: Bearer <YOUR_API_KEY>' \
--data-raw '{
  "provider": "openrouter",
  "model": "gemini-pro-1.5",
  "max_tokens": 150,
  "messages": [
    {
      "role": "user",
      "content": "What are the key differences between machine learning and deep learning?"
    }
  ]
}'

Response Example​

The expected response structure for a Google model might look like this:

{
  "id": "chatcmpl-12345example12345",
  "object": "chat.completion",
  "created": 1729535643,
  "provider": "openrouter",
  "model": "gemini-pro-1.5",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Machine learning and deep learning differ in several key aspects:\n\n1. **Complexity**: Machine learning uses simpler algorithms for pattern recognition, while deep learning uses complex neural networks with multiple layers.\n\n2. **Data Requirements**: Machine learning can work with smaller datasets, but deep learning typically needs vast amounts of data to be effective.\n\n3. **Feature Extraction**: Machine learning often requires manual feature engineering, while deep learning automatically learns and extracts relevant features.\n\n4. **Hardware Requirements**: Machine learning algorithms can run on standard computers, but deep learning usually needs powerful GPUs for efficient processing.\n\n5. **Applications**: Machine learning is suited for structured data analysis, while deep learning excels in complex tasks like image recognition and natural language processing."
      },
      "logprobs": null,
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 18,
    "completion_tokens": 132,
    "total_tokens": 150,
    "prompt_characters": 72,
    "response_characters": 548,
    "cost": 0.003000,
    "latency_ms": 2800
  },
  "system_fingerprint": "fp_123abc456def"
}

API Highlights​

  • Provider: Specify the provider or leave blank for automatic selection.
  • Model: Use any model from the Google Series, such as gemini-pro-1.5 or others suited to your task. See Models Guide.
  • Max Tokens: Set the maximum response token count (e.g., 150 in this example).
  • Messages: Format your request with a sequence of messages, including user input and system instructions. See message formatting.

This example demonstrates how to seamlessly query models from the Google Series for conversational or instructional tasks.

Applications and Integrations​

Ethical Considerations​

Google's AI models are designed with responsible AI principles in mind. Users should implement appropriate safeguards and consider potential biases in model outputs. For guidance on responsible AI usage, see Google's AI Principles, Responsible AI Practices.

Licensing​

The Google Series includes both proprietary and open-source models. While Gemini models are available through Google Cloud with specific terms of service, Gemma models are released under the Apache 2.0 license for research and commercial use. For detailed licensing information, consult the model-specific documentation, usage guidelines.

tip

Try out the Google models in APIpie's various supported integrations.