This model is deprecated. We still benchmark its performance, but less often: every 18 hours instead of every 6.
Google has launched a newer release, Gemma 3n E4B Instruct. We suggest considering it instead.
Gemma 3n E4B Instruct Preview (May '25): Intelligence, Performance & Price Analysis
Model summary
Performance
Verbosity
Gemma 3n E4B Instruct Preview (May '25) supports text, speech, and video input, outputs text, and has a 32k tokens context window with knowledge up to August 2024.
| Reasoning | No |
|---|---|
| Input modality | Supports: text, speech, and video |
| Output modality | Supports: text |
| Knowledge cutoff | Aug 1, 2024 |
| Context window | 32k ~48 A4 pages of size 12 Arial font |
| Total parameters | 8.4B |
| Active parameters | 4B Number of parameters active per token during inference |
| License | Gemma license |
| Model weights | Hugging Face |
Metrics are compared against models of the same class:
- Non-reasoning models → compared only with other non-reasoning models
- Reasoning models → compared across both reasoning and non-reasoning
- Open weights models → compared only with other open weights models of the same size class:
- Tiny: ≤4B parameters
- Small: 4B–40B parameters
- Medium: 40B–150B parameters
- Large: >150B parameters
- Proprietary models → compared across proprietary and open weights models of the same price range, using a blended 3:1 input/output price ratio:
- <$0.15 per 1M tokens
- $0.15–$1 per 1M tokens
- >$1 per 1M tokens
Time per Task
Intelligence
Artificial Analysis Intelligence Index
Artificial Analysis Intelligence Index by Open Weights / Proprietary
Intelligence Index vs. Cost per Task
Openness Index
Artificial Analysis Openness Index: Score
Cost
Cost per Intelligence Index Task
Cost to Run Artificial Analysis Intelligence Index
Pricing: Cache Hit, Input, and Output
Context Window
Context Window
Model Size (Open Weights Models Only)
Model Size: Total and Active Parameters
Further details
Weights | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Gemma 3n E4B Instruct Preview (May '25) | - | 8.4B 4B active at inference time | 32k | - | - | - | - |
Frequently Asked Questions
Common questions about Gemma 3n E4B Instruct Preview (May '25)
Gemma 3n E4B Instruct Preview (May '25) was released on May 20, 2025.
Gemma 3n E4B Instruct Preview (May '25) was created by Google.
No, Gemma 3n E4B Instruct Preview (May '25) is not a reasoning model. It provides direct responses without extended chain-of-thought reasoning.
Gemma 3n E4B Instruct Preview (May '25) supports text, speech, and video input.
Gemma 3n E4B Instruct Preview (May '25) supports text output.
No, Gemma 3n E4B Instruct Preview (May '25) does not support image input. It can only process text, speech, and video.
Yes, Gemma 3n E4B Instruct Preview (May '25) is multimodal. It can process text, speech, and video input and generate text output.
Gemma 3n E4B Instruct Preview (May '25) has a context window of 32k tokens. This determines how much text and conversation history the model can process in a single request.
Yes, Gemma 3n E4B Instruct Preview (May '25) is open weights. The model weights are publicly available and can be downloaded for self-hosting.
Gemma 3n E4B Instruct Preview (May '25) has 8.4B parameters (4B active).
Gemma 3n E4B Instruct Preview (May '25) is a Mixture of Experts (MoE) model with 8.4B total parameters, but only 4B active parameters are used during inference.
Gemma 3n E4B Instruct Preview (May '25) is released under the Gemma license license. This license allows commercial use. View license
Gemma 3n E4B Instruct Preview (May '25) has a knowledge cutoff of August 2024. The model's training data includes information up to this date.
Gemma 3n E4B Instruct Preview (May '25) is an open weights model that can be self-hosted. View providers
Gemma 3n E4B Instruct Preview (May '25) is an open weights model that can be downloaded and self-hosted. Compare providers