Google makes EmbeddingGemma 300M, an open-source embedding model. It takes in text and outputs vector representations.
Input
Output
Context
2K
Max Output
n/a
Parameters
302.9M
Input Modalities
Output Modalities
None the catalog can show from stated facts or measured scores.
Download 1.21 GB · Runs on vLLM v0.32.0 · 34 GGUF builds
| Precision | Weights | Smallest GPU that fits |
|---|---|---|
| FP16 / BF16 | 0.56 GB | 1x NVIDIA T4 (16 GB) |
| INT8 | 0.28 GB | 1x NVIDIA T4 (16 GB) |
| INT4 | 0.14 GB | 1x NVIDIA T4 (16 GB) |
Computed by Inferbase from the 302,863,104 parameters in the creator's Hugging Face repository; weights only, as serving adds the framework and the input being processed.
Specification sources
As of Nov 14, 2026.
Answered from the facts on this page, with their sources.
EmbeddingGemma 300M has a context window of 2,048 tokens (source: Creator or provider API). A provider may serve less; Inferbase routes on what each provider serves.
Yes. EmbeddingGemma 300M is an open-weight model under the Gemma license. You can download and self-host it within the license terms.
EmbeddingGemma 300M has 302,863,104 parameters (302.9M), as stated in the creator's Hugging Face repository.
EmbeddingGemma 300M accepts Text and produces Embedding. The Capabilities card lists the API features it supports, such as function calling and structured output.
No benchmark source has published a result for EmbeddingGemma 300M yet, so it is shown as unmeasured, not estimated.
Not on Inferbase Managed today. Add your own provider key or endpoint that serves EmbeddingGemma 300M, and Inferbase runs it on your account with the same routing and logs.
The weights take 0.6 GB at FP16 / BF16, 0.3 GB at INT8, 0.1 GB at INT4, computed from the 302,863,104 parameters its creator publishes. Serving adds the framework and the input being processed.
Start on the free plan with 5,000 requests a month, no card required.