Gemini Embedding Models by Google DeepMind
Google DeepMindProprietary
5 models2023–2026Up to 2K ctxFrom $0.1/1M input
About
Google's Gemini Embedding models for generating text and multimodal embeddings. Used for semantic search, retrieval, classification, and clustering tasks.
Current Variants
Use-when guidance is derived from seed capabilities, context, release, and replacement fields.
5 in view
Gemini Embedding 2 PreviewCurrent
Use when the workload needs embedding and multimodal inputs.
2026-04embeddingmultimodal inputs
Gemini Embedding 2Current
Use when the workload needs embedding, multimodal inputs, and audio.
2026-04embeddingmultimodal inputsaudio
Multimodal EmbeddingsCurrent
Use when the workload needs embedding and multimodal inputs.
2024-08embeddingmultimodal inputs
Use when the workload needs 2K context.
2023-022K context
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Gemini Embedding 2 Preview | Use when the workload needs embedding and multimodal inputs. | 2026-04 | embeddingmultimodal inputs | Current |
| Gemini Embedding 2 | Use when the workload needs embedding, multimodal inputs, and audio. | 2026-04 | embeddingmultimodal inputsaudio | Current |
| Multimodal Embeddings | Use when the workload needs embedding and multimodal inputs. | 2024-08 | embeddingmultimodal inputs | Current |
| Gemini Embedding | Use when the workload needs embedding. | 2023-12 | embedding | Current |
| text-embedding-004 on Google Vertex AI | Use when the workload needs 2K context. | 2023-02 | 2K context | Current |
Release Timeline
4 release groups2026-04
2 current
Gemini Embedding 2
Currentembeddingmultimodal inputsaudio
Gemini Embedding 2 Preview
Currentembeddingmultimodal inputs
2024-08
1 current
Multimodal Embeddings
Currentembeddingmultimodal inputs
2023-12
1 current
Gemini Embedding
Currentembedding
2023-02
1 current
text-embedding-004 on Google Vertex AI
Current2K context
Specifications(5 models)
| Model | Released | Context | Vision | Multimodal |
|---|---|---|---|---|
| Gemini Embedding 2 Preview | 2026-04 | — | Yes | Yes |
| Gemini Embedding 2 | 2026-04 | — | Yes | Yes |
| Multimodal Embeddings | 2024-08 | — | Yes | Yes |
| Gemini Embedding | 2023-12 | — | No | No |
| text-embedding-004 on Google Vertex AI | 2023-02 | 2K | No | No |
Available From(2 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| text-embedding-004 on Google Vertex AI | GCP Vertex AI | $0.1 | — | Serverless |
| Gemini Embedding | Google AI Studio | $0.15 | — | Serverless |
| Gemini Embedding | GCP Vertex AI | $0.15 | — | Serverless |
| Gemini Embedding 2 Preview | Google AI Studio | $0.2 | — | Serverless |
| Gemini Embedding 2 | Google AI Studio | $0.2 | — | Serverless |
Frequently Asked Questions
- What is Gemini Embedding used for?
- Gemini Embedding is used for embedding and vision and multimodal work. The family description and listed model capabilities point to those workloads as the best fit.
- How does Gemini Embedding compare to Gemma 4?
- Gemini Embedding by Google DeepMind is strongest where you need embedding, while Gemma 4 by Google DeepMind is the closest related family to check for vision and multimodal work. Gemini Embedding has 5 listed variants and reaches up to 2K context, while Gemma 4 reaches up to 256K context, so compare the specs and pricing tables before choosing a production model.
- Which Gemini Embedding model should I use?
- For the lowest listed input price, start with text-embedding-004 on Google Vertex AI through GCP Vertex AI at $0.1/1M input tokens. For the most capable/latest local choice, evaluate Gemini Embedding 2 Preview with multimodal inputs.
