Strengths
- Only 3.8 billion of its 25.2 billion parameters work on each token, so it generates faster than a dense model of similar quality.
- It comes close to the larger Gemma 4 31B on many tests: GPQA Diamond 82.3% vs 84.3%, AIME 2026 88.3% vs 89.2%.
- The Apache 2.0 license allows commercial use and changes to the model.
Weaknesses
- It does not save memory: all parameters must be loaded on the GPU, about 14.4 GB in 4-bit and 28.8 GB in 8-bit.
- Clearly weaker than 31B on agentic and hard tasks: Tau2 68.2% vs 76.9%, Humanity's Last Exam 8.7% vs 19.5%, long-context MRCR v2 44.1% vs 66.4%.
- Its knowledge stops in January 2025.
Best for
- fast local use on a single GPU
- private chatbots and search over company documents (RAG)
- self-hosted workloads with no per-token fees
Not ideal for
- retrieval across long documents
- complex agentic tasks, where Gemma 4 31B is better
How to use it in Europe
You can download the weights for free from Hugging Face or Kaggle and run them in apps such as Ollama or LM Studio, or on your own server. The Apache 2.0 license has no geographic limits, so the model can run inside your company in the EU and your data stays in house. In the Gemini API and Google AI Studio, Gemma 4 is only on the free tier, with no paid tier for it.
Benchmark results
| Benchmark | Score | Source |
|---|---|---|
| MMLU Pro | 82.6% | ai.google.dev ↗, April 2026 |
| GPQA Diamond | 82.3% | ai.google.dev ↗, April 2026 |
| LiveCodeBench v6 | 77.1% | ai.google.dev ↗, April 2026 |
| AIME 2026 (no tools) | 88.3% | ai.google.dev ↗, April 2026 |
Benchmarks cover only a narrow slice of ability and vendors often report them at their best settings. Treat them as a rough guide.
More models from this vendor
- Gemini 4 Carbon: Google's most capable model, announced on 28 October 2026. So far only selected cybersecurity experts have access.
- Gemini 4 Argon: Google's most capable publicly available model, out in October 2026. Built for demanding analytical work where accuracy matters and made-up answers are costly.
- Gemini 3.8 Flash: Google's fast, lower-cost model for everyday work, coding and document analysis. You can try it in the API for free.
- Gemini 3.5 Flash-Lite: A very fast, low-cost Google model for large volumes of simple tasks, such as pulling data out of documents.
- Gemma 4 31B: Google's largest open model, which you can run on your own computer or server when data must not leave the company.
Similar models from other vendors
- Muse Glimmer 30B (Meta): Meta's small open model under Apache 2.0 that runs on a single graphics card, even offline.
- Llama 4 Scout (Meta): The smaller Llama 4 from April 2025, very cheap through providers. Its license grants no rights to people or companies in the EU.
- Ministral 3 14B (Mistral AI): Mistral's small open model under Apache 2.0 that runs on a single GPU, even without an internet connection.
- Nemotron 3.5 Bolt 74B-A9B (NVIDIA): NVIDIA's open model for AI agents on your own powerful PC, small enough to fit in 24 GB of graphics card memory.
- GPT-6 Luna (OpenAI): OpenAI's cheapest model, made for high volumes of simple tasks, sorting text and low-cost chatbots.
- gpt-oss-20b (OpenAI): OpenAI's smaller open model that runs on a computer with 16 GB of memory, even without internet access.