# Gemma 4 E2B

Model tracker · checked 2026-11-05 · https://dbggr.dev/models/gemma-4-e2b/

Small open-weight model (2.3B effective) for low-latency text, image and audio tasks

Gemma 4 E2B is a fast open-weight model from Google, released 31 March 2026. At AWS it costs $0.04 per million input tokens and $0.08 per million output tokens, with a 128K-token context window.

- Made by: Alphabet (Google, Google DeepMind, Google Cloud)
- Released: 2026-03-31
- Status: Generally available
- Tier: Fast
- Weights: Open
- Memory to run: About 13 GB (BF16)
- Context window: 128,000 tokens
- Input: text, image, audio, video
- API id: google.gemma-4-e2b
- Docs: https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-google-gemma-4-e2b.html

## Gemma 4 E2B price at AWS

US dollars per million tokens. Blended is 3 input tokens for every output token, one number to compare on.

| Provider | Input | Output | Cached input | Blended | Batch | Checked |
|---|---|---|---|---|---|---|
| AWS | $0.04 | $0.08 | – | $0.05 | – | [2026-11-05](https://aws.amazon.com/bedrock/pricing/) |

- AWS: Amazon Bedrock, US East (N. Virginia), Standard tier, bedrock-mantle endpoint only. Same in US East (Ohio) and US West (Oregon); Europe (Frankfurt): $0.05 / $0.10.

## Can you run Gemma 4 E2B yourself?

Yes. Google publishes the weights on Hugging Face as [google/gemma-4-E2B-it](https://huggingface.co/google/gemma-4-E2B-it), under Apache 2.0: 5.1B (2.3B effective) parameters, 10 GB in BF16.

Running it takes about 13 GB of GPU memory: a single GPU with at least 13 GB of memory, or a Mac with comfortably more than 13 GB. That's the weight files plus 20% for working memory with short inputs; long contexts need more, and fitting in memory doesn't make it fast.

## Models like Gemma 4 E2B

| Model | Input | Output | Context | Weights |
|---|---|---|---|---|
| Gemma 4 E2B | $0.04 | $0.08 | 128K | Open |
| [gpt-oss-20b](https://dbggr.dev/models/gpt-oss-20b/) | $0.03 | $0.14 | 131K | Open |
| [gpt-oss-120b](https://dbggr.dev/models/gpt-oss-120b/) | $0.037 | $0.17 | 131K | Open |
| [NVIDIA Nemotron 3.5 Lightning 30B A3B](https://dbggr.dev/models/nvidia-nemotron-3-5-lightning-30b-a3b/) | $0.06 | $0.16 | 262K | Open |
| [Ministral 3 3B](https://dbggr.dev/models/ministral-3-3b/) | $0.10 | $0.10 | 256K | Open |
| [LFM3-1.7B](https://dbggr.dev/models/lfm3-1-7b/) | $0.08 | $0.20 | 262K | Open |

## FAQ

### How much does Gemma 4 E2B cost?

Gemma 4 E2B costs $0.04 per million input tokens and $0.08 per million output tokens at AWS. Prices are list prices in US dollars, last checked 5 November 2026.

### What is Gemma 4 E2B's context window?

Gemma 4 E2B takes up to 128,000 tokens in one request.

### Is Gemma 4 E2B open weight?

Yes. Google publishes the weights, so you can run Gemma 4 E2B yourself or at any host that serves it.

### Can I run Gemma 4 E2B locally?

Gemma 4 E2B's weights are 10 GB in BF16, so it needs about 13 GB of GPU memory: a single GPU with at least 13 GB of memory, or a Mac with comfortably more than 13 GB.

### Where can I use Gemma 4 E2B?

Gemma 4 E2B is available from AWS. At AWS, call it as `google.gemma-4-e2b`.