Google Gemini¶

Moved to Separate Package

As of v0.15.0, Gemini has been moved to a separate package to reduce dependencies in omnillm-core.

Install the Gemini provider:

go get github.com/plexusone/omnillm-gemini

Overview¶

Models: Gemini-2.5-Pro, Gemini-2.5-Flash, Gemini-1.5-Pro, Gemini-1.5-Flash
Features: Chat completions, streaming, massive context windows
SDK: Uses official google.golang.org/genai SDK

Installation¶

go get github.com/plexusone/omnillm-gemini

Configuration¶

import (
    "github.com/plexusone/omnillm-core"
    _ "github.com/plexusone/omnillm-gemini" // Register Gemini provider
)

client, err := omnillm.NewClient(omnillm.ClientConfig{
    Providers: []omnillm.ProviderConfig{
        {Provider: omnillm.ProviderNameGemini, APIKey: "your-gemini-api-key"},
    },
})

Available Models¶

Model	Context Window	Description
`omnillm.ModelGemini25Pro`	1M	Gemini 2.5 Pro
`omnillm.ModelGemini25Flash`	1M	Gemini 2.5 Flash (fast)
`omnillm.ModelGemini15Pro`	2M	Gemini 1.5 Pro (largest context)
`omnillm.ModelGemini15Flash`	1M	Gemini 1.5 Flash

JSON Mode¶

Gemini supports JSON mode for structured outputs:

response, err := client.CreateChatCompletion(ctx, &omnillm.ChatCompletionRequest{
    Model: omnillm.ModelGemini25Pro,
    Messages: messages,
    ResponseFormat: &omnillm.ResponseFormat{Type: "json_object"},
})

Large Context¶

Gemini 1.5 Pro supports up to 2 million tokens of context, making it ideal for:

Long document analysis
Large codebase understanding
Extended conversation history