Gemini Embedding 2
Gemini Embedding 2 by Google AI: $0.20 per 1M input and — per 1M output tokens, 8K context. Call it with one OpenAI-compatible API key, compare it with similar models, start free.
Gemini Embedding 2 is served by Google AI and called through odnoga on /v1/embeddings, in the same OpenAI-compatible shape every other model on that endpoint uses. It accepts images alongside text. Its 8K-token context window sets how much input you can send in one request.
Specification and price
| Vendor | Google AI |
|---|---|
| Model ID | gemini-embedding-2 |
| Context window | 8K tokens |
| Max output | 1 tokens |
| Served on | /v1/embeddings |
| Capabilities | Reads images |
| Per 1M tokens | Vendor cost | Your price on Free+7% |
|---|---|---|
| Input | $0.2 | $0.215 |
| Output | — | — |
| Cached input | — | — |
Vendor cost is the list price per million tokens as recorded in the odnoga catalog; your price applies your plan margin with the same formula that bills every request — see pricing. Pricing
Call it through odnoga
const res = await fetch("https://api.odnoga.com/v1/embeddings", {
method: "POST",
headers: {
Authorization: `Bearer ${process.env.ODNOGA_API_KEY}`,
"Content-Type": "application/json",
},
body: JSON.stringify({
model: "gemini-embedding-2",
input: "The text you want to embed",
}),
});
One endpoint, one key, one bill — /v1/embeddings takes the same shape whichever vendor is behind it. Swap the model id and odnoga handles keys, routing, limits and cost accounting.
What Gemini Embedding 2 can do
Everything below is what the model itself supports. Ask for something it does not do and odnoga tells you before the request reaches the provider, instead of returning a worse answer.
Send a screenshot, a photo or a scan in the same message as your text.
Beyond text
Takes recorded speech as input.
Takes video as input and answers questions about what happens in it.
How to use Gemini Embedding 2
Derived from the odnoga catalog record for this model.
Best for
- Turning text into vectors for search, clustering, deduplication and retrieval — it does not answer questions, it measures similarity.
Not the right pick when
- Agent loops that must call your functions — tool calling is not available here.
- Pipelines that require guaranteed JSON — parse defensively or pick a model with enforced JSON.
- Very long generated documents — output is capped at 1 tokens per response.
- Large documents in one request — the window is 8K tokens, so you will need chunking.
- Anything where a wrong answer is costly without a human check — no model in the catalog removes that requirement.
Practical tips through odnoga
- 01Pin the model id in a managed prompt version, so a model swap is a version change you can compare and roll back, not an edit in application code.
- 02Compare it against 2–8 other models on the same frozen test cases in the evaluation laboratory before you make it the production default.
- 03For repeated identical deterministic calls, odnoga answer reuse returns the stored answer and bills no vendor tokens — turn it off for creative output.
- 04Set a fallback model on the route so a vendor incident degrades quality instead of returning an error, and a per-tenant budget so one caller cannot spend the month.
If Gemini Embedding 2 is not the one
Models that take the same kind of input, are priced the same way, and do the closest thing to what this one does. Not a quality ranking — odnoga does not benchmark models.
| Model | Context | Input / 1M | Output / 1M | Capabilities |
|---|---|---|---|---|
| Gemini Embedding 2 | 8K | $0.2 | — | Reads images |
| Text Embedding 3 Large | 8K | $0.13 | — | Batch |
| Text Embedding Ada 002 | 8K | $0.1 | — | Batch |
| Text Embedding 3 Small | 8K | $0.02 | — | Batch |
Questions
- How much does Gemini Embedding 2 cost per 1M tokens?
- Google AI lists $0.2 / — per 1M input / output tokens in the odnoga catalog. Through odnoga you pay that vendor price plus your plan margin, and every request is recorded with both numbers.
- What is Gemini Embedding 2 best for?
- Turning text into vectors for search, clustering, deduplication and retrieval — it does not answer questions, it measures similarity.
- Can I switch to Gemini Embedding 2 without changing my code?
- Yes. odnoga exposes one OpenAI-compatible endpoint, so switching means sending "gemini-embedding-2" as the model id — or changing it in the managed prompt version, with no application deploy.
- How large is the Gemini Embedding 2 context window?
- 8K tokens of input, with up to 1 tokens of output per response.
All models · Pricing · Docs · Compare models in the evaluation lab
Where this model sits
Other Google AI models
Gemini 2.5 Flash Native Audio (Live)
$0.5 / $2 per 1M
Gemini 2.5 Flash TTS
$0.5 / $10 per 1M
Gemini 2.5 Pro TTS
$1 / $20 per 1M
Gemini 3 Flash Preview
$0.5 / $3 per 1M
Gemini 3 Pro Image (Nano Banana Pro)
$2 / $12 per 1M
Gemini 3.1 Flash Image (Nano Banana 2)
$0.5 / $3 per 1M
Gemini 3.1 Flash Lite Image (Nano Banana 2 Lite)
$0.25 / $1.50 per 1M
Gemini 3.1 Flash Live
$0.75 / $4.50 per 1M