Gemma 4 E4B
Gemma 4 E4B is an on-device Gemma 4 model with about 4.5B effective parameters (8B including per-layer embeddings). It accepts text, images and audio, with video handled as frames, and supports a 128K context window. Google also publishes quantization-aware and mobile builds.
Specifications
- Context window
- 128Kabout 98,000 words, or 200 pages
- Max output
- Not published
- Input price
- —no list price confirmed
- Output price
- —
- Cached input
- Not published
- Weights
- OpenApache-2.0
- Parameters
- 8Bdense
- Input types
- Text, Images, Audio, Video
- Output types
- Text
- Reasoning mode
- Yes
- Tool calling
- Yes
- Released
- Mar 31, 2026Current
Not listed on the official pricing pages of Together AI, Fireworks, DeepInfra or Groq as of 2026-10-10; intended mainly for local and on-device use. Google quotes 4.5B 'effective' parameters; the 8B total includes per-layer embeddings.
What it costs in practice
| Workload | Monthly cost at list price |
|---|---|
| Support chatbot30,000 conversations a month, 2,000 input and 400 output tokens each | No list price |
| Document summaries2,000 documents a month, 12,000 input and 800 output tokens each | No list price |
| Coding agent1,000 tasks a month, 60,000 input and 6,000 output tokens each | No list price |
Standard rates, no caching or batch discounts. Run your own numbers.
Strengths
- Audio, image and text input in a phone-class model
- Per-layer embeddings keep the compute footprint near 4.5B
- Apache 2.0 with official QAT and mobile builds
Trade-offs to weigh
- 128K context is half that of the larger Gemma 4 models
- Small capacity limits complex reasoning
Where it fits
Google positions Gemma 4 E4B for:
Where you can use it
- Hugging Face
- Ollama
Run it yourself
Released under Apache-2.0. Approximate memory needed just to load the weights:
| Precision | Memory for weights |
|---|---|
| 16-bit (BF16/FP16) | 16 GB |
| 8-bit (FP8/INT8) | 8 GB |
| 4-bit (Q4_K_M, about 4.85 bits) | 5 GB |
Add memory for the KV cache (it grows with context length and concurrent users) and 1-2 GB of runtime overhead. Size it precisely in the VRAM calculator.
Alternatives to consider
| Model | Tier | Context | Input / output per 1M | |
|---|---|---|---|---|
| Ministral 3 8BMistral AI | Compact | 256K | $0.15 / $0.15 | Compare |
| Ministral 3 3BMistral AI | Compact | 256K | $0.10 / $0.10 | Compare |
| Qwen3.6-35B-A3BAlibaba (Qwen) | Compact | 256K | $0.10 / $0.95 | Compare |
| gpt-oss-20bOpenAI | Compact | 128K | $0.075 / $0.30 | Compare |
Frequently asked questions
How much does Gemma 4 E4B cost?
We couldn't confirm a published per-token list price for Gemma 4 E4B on October 10, 2026. Not listed on the official pricing pages of Together AI, Fireworks, DeepInfra or Groq as of 2026-10-10; intended mainly for local and on-device use. Google quotes 4.5B 'effective' parameters; the 8B total includes per-layer embeddings. Check the provider's pricing page for current rates.
What is the context window of Gemma 4 E4B?
Gemma 4 E4B accepts up to 131,072 tokens of context (about 98,000 words, or 200 pages). Long prompts cost more and can be slower, so most apps send only what the task needs.
Is Gemma 4 E4B open source?
Gemma 4 E4B is an open-weight model released under the Apache-2.0 license. You can download the weights from Hugging Face (google/gemma-4-E4B-it) and run or fine-tune them yourself, within the license terms. Open weights isn't the same as fully open source: training data and code may not be published.
Can I run Gemma 4 E4B locally?
Yes, if you have enough memory. The weights alone need about 5 GB at 4-bit quantization and about 16 GB at 16-bit, plus room for the KV cache. Use the VRAM calculator on LLModel.com to check a specific GPU or Mac.
What is Gemma 4 E4B best for?
Google positions Gemma 4 E4B for on-device, low latency, multimodal, self-hosting. It accepts text, images, audio, video as input, has a reasoning (thinking) mode and supports tool calling. Test it on a sample of your own tasks before you choose it for production.
Sources
- huggingface.co/google/gemma-4-E4B-it
- ai.google.dev/gemma/docs/releases
- huggingface.co/google/gemma-4-31B-it
- huggingface.co/google/gemma-4-E4B-it/raw/main/config.json
- ollama.com/library
Facts checked against these pages on October 10, 2026. Spot an error? Tell us.
Still choosing a model?
Send us your use case and constraints. We'll come back with a shortlist and an evaluation plan you can run on your own data.
- Free and vendor-neutral: no model provider pays for placement.
- Your details go only to partners you agree to.