Google's recommended embedding model — first multimodal embedding model in the Gemini API (text-only via this endpoint today). Default output dimension 3072; supports the `dimensions` parameter (recommended: 768, 1536, 3072). Embedding space is incompatible with gemini-embedding-001.
gemini-embedding-2Compare price, limits and live performance side by side. LLM Gateway routes each request to the best provider that can handle your prompt size and parameters.
1 provider
| Provider | Actions | |||||
|---|---|---|---|---|---|---|
| 8.2k | $0.20 |
Prices are per provider at standard rates; open a row for regions, tiers and cache pricing. Throughput, latency and uptime come from LLM Gateway traffic over the last 24 hours and are hidden under 1,000 requests.
Google's recommended embedding model — first multimodal embedding model in the Gemini API (text-only via this endpoint today). Default output dimension 3072; supports the `dimensions` parameter (recommended: 768, 1536, 3072). Embedding space is incompatible with gemini-embedding-001. You can access it through LLM Gateway's OpenAI-compatible API with automatic provider routing, fallback, and cost analytics.
Pricing for Gemini Embedding 2 on LLM Gateway starts at $0.20 per million input tokens and $0.00 per million output tokens, depending on the provider. The pricing table above always reflects the current per-provider rates.
Gemini Embedding 2 supports a context window of up to 8,192 tokens on its largest provider deployment.
Gemini Embedding 2 is served by Google AI Studio through LLM Gateway. Requests are automatically routed to the best available provider, with fallback when a provider has issues.
No. Gemini Embedding 2 does not currently support tool calling or structured JSON outputs through LLM Gateway.
Gemini Embedding 2 was released on April 22, 2026.