兔子API 兔子API
中

model

Google Gemini

Direct answerGemini is Google's multimodal family; the official catalog lists separate models and methods for general reasoning, low latency, images, video, audio, embeddings, and agents.

Updated · Reviewed

Official portfolio

The Gemini catalog includes Gemini 3 general models as well as specialized image, Veo video, Lyria music, embedding, Computer Use, and Deep Research offerings. They share a brand but not a single request protocol or output modality.

Multimodal capability

General Gemini models combine text, image, audio, and video through contents and parts; specialized models use their own generation methods. The model page's supported generation methods are the authority for functions, caching, and structured output.

Choosing within the family

Evaluate capability-first models for advanced reasoning and multimodal understanding, Flash or Lite classes for interactive and high-volume work, and dedicated models for image, video, music, or vectors. Media pricing and limits are not directly comparable to text-token models.

API use

Native calls use /v1beta/models/{model}:generateContent or a task-specific method. Enabled models on this site use its Bearer token; an OpenAI-compatible route may not express every Gemini part, safety option, or grounding field.

Lifecycle

The official catalog marks preview, experimental, deprecated, and shut-down states. Avoid retiring IDs in production, pin a stable identifier, and verify region, file limits, context, output, safety settings, and data policy.

Use cases

  • Native multimodality
  • Long-context reasoning
  • Media generation

API protocols

  • /v1/chat/completions

FAQ

Is the model name permanent?

No. Official aliases can move and enabled API IDs can change; pin a tested ID in production and verify availability separately in the model marketplace.

How do I choose between general Gemini and image, video, music, or embedding models?

Use general capability models for advanced reasoning and multimodal understanding and evaluate Flash or Lite classes for interactive volume. Images, video, music, and vectors use dedicated models and generation methods; a shared Gemini brand does not imply the same endpoint, output, or billing unit.

Is native Gemini GenerateContent equivalent to OpenAI-compatible access?

No. Native contents/parts can express multimodal content and Gemini-specific safety or grounding fields that a compatibility layer may not preserve. Choose from the model page's supported generation methods and run field-level tests.

  • Google GeminiGoogle recommends Interactions API for new direct integrations. This gateway currently exposes Gemini-native GenerateContent and streamGenerateContent, not Interactions, and recommends Bearer while accepting x-goog-api-key compatibility.
  • Gemini API guide: Interactions and GenerateContentGoogle recommends Interactions API for new projects; this gateway currently exposes Gemini-native compatibility mainly through GenerateContent, and the two request and stream formats are not interchangeable.
  • Image generation and vision model guideVision understanding and image generation use different interfaces; generation projects must validate prompt adherence, editing consistency, text, likeness rights, moderation, and output terms.
  • Tool use and AI agent model guideAn agent is a controlled loop, not merely tool calling: structured schemas, least privilege, state, idempotency, timeouts, approval, audit, and replayable evaluation.

Official sources

  1. Gemini Models Official