Gemma 3 4B
Gemma 3 4B (Google DeepMind) was released on Mar 12, 2025. In the Epoch AI capability ranking (ECI), it ranks #210 out of 267 models, with 116.01.
Where Gemma 3 4B stands out and where it falls behind, by the numbers
Gemma 3 4B has published results on 4 tests compiled by Epoch AI. The yardstick here is its generation: its position among the models released between Jun 15, 2024 and Jun 10, 2025 that took the same test. Against the whole archive, which includes 2023 models, a recent release lands near the top of almost everything, and that says nothing.
Where it stands out
Tests where it sits in the top third of its own generation.
It is not in the top third of its generation on any test with 6 or more models compared.
Where it falls behind
Tests where it sits in the bottom half of its own generation.
See all 4 tests, one by one
| Test | Score | In its generation | Overall | Test record |
|---|---|---|---|---|
| GPQA Diamond | 23.2% | #55 / 55 | #171 / 173 | 96.4% Gemini 4 Argon |
| AIME 2024 and 2025 (OTIS) | 7.5% | #39 / 52 | #139 / 168 | 100.0% GPT-5.5 |
| Dtbench | 50.9% | #35 / 40 (tied with 1) | #134 / 148 | 98.9% Claude Opus 5.5 |
| Lmca | 2.78 | #29 / 29 | #121 / 121 | 68.23 Claude Opus 5.5 |
For each test, the model's best effort level counts, as in the ranking. Source of the scores: Epoch AI, under CC BY 4.0.
What to compare Gemma 3 4B with
The models from other labs closest in capability and released in the same period, at most two per lab. Click any of them to see its page.
| Model | Lab | ECI | ELO | LiveBench | Release |
|---|---|---|---|---|---|
| Mistral NeMo | Mistral AI | 118.71 | n/a | n/a | Jul 18, 2024 |
| Qwen2.5-7B | Alibaba | 118.53 | n/a | n/a | Sep 19, 2024 |
| Ministral 3B | Mistral AI | 118.12 | n/a | n/a | Oct 16, 2024 |
| Llama 3.1-8B | Meta AI | 116.61 | n/a | n/a | Jul 23, 2024 |
| Qwen2.5-Coder-14B | 116.44 | n/a | n/a | Sep 18, 2024 | |
| Gemma 3 4B | Google DeepMind | 116.01 | n/a | n/a | Mar 12, 2025 |
| Qwen2.5-Coder (7B) | Alibaba | 113.20 | n/a | n/a | Sep 18, 2024 |
Other models from Google DeepMind
The models from the same lab closest in time, newest first.
| Model | Release | ECI | ELO | LiveBench | Access |
|---|---|---|---|---|---|
| Gemini 2.5 Pro (Jun 2025) | Jun 5, 2025 | 145.31 | 1,507 | n/a | API access |
| Gemini 2.5 Pro Exp (Mar 2025) | Mar 25, 2025 | 144.20 | n/a | n/a | API access |
| Gemma 3 27B | Mar 12, 2025 | 130.05 | n/a | n/a | Open weights (restricted use) |
| Gemma 3 12B | Mar 12, 2025 | 123.49 | n/a | n/a | Open weights (restricted use) |
| Gemma 3 4B | Mar 12, 2025 | 116.01 | n/a | n/a | Open weights (restricted use) |
| Gemini 2.0 Pro Exp (Feb 2025) | Feb 5, 2025 | 135.08 | n/a | n/a | Hosted access (no API) |
| Gemini 2.0 Pro | Feb 5, 2025 | 135.05 | n/a | n/a | Hosted access (no API) |
| Gemini 2.0 Flash Thinking Exp | Jan 21, 2025 | 135.39 | n/a | n/a | API access |
| Gemini 2.0 Flash Thinking (Jan 2025) | Jan 21, 2025 | 135.36 | n/a | n/a | API access |
Have you used Gemma 3 4B? Rate it
Users' opinion, not measurement: the rating is kept apart from the ranking and does not move anyone's position. One vote per browser, no sign-up, and you can change your rating later.
What people ask about Gemma 3 4B
Where does Gemma 3 4B rank among AI models?
In the ROO3 AI Benchmark, ordered by the Epoch AI ECI, Gemma 3 4B ranks #210 out of 267 models, with an ECI of 116.01. The ECI is a composite index that combines several hard benchmarks into a single scale. The position changes when a new model comes out, and the page is refreshed six times a day.
Where do the numbers on this page come from?
From the same three sources as the ranking, all openly licensed: Epoch AI (ECI and individual tests, CC BY 4.0), Arena (human-preference ELO, CC BY 4.0) and LiveBench (overall average, CC BY-SA 4.0). ROO3 does not measure any model. When the page shows a figure published by the maker itself, it appears in the written analysis, labeled as the maker's, and never enters the ranking.
Where each number comes from
Epoch AI, "AI Benchmarking Hub". Published at epoch.ai. Data under the CC BY 4.0 licence, reordered by roo3.co.
The Arena (LMArena) leaderboard, from the public lmarena-ai/leaderboard-dataset on Hugging Face, under the CC BY 4.0 licence. We are not affiliated with the Arena.
LiveBench, by the LiveBench team, published at livebench.ai under the CC BY-SA 4.0 license. The overall average is recomputed by roo3.co with the same formula as the website (average of the categories), and those numbers stay under CC BY-SA 4.0. We are not affiliated with LiveBench.
← See the full ranking