Updated daily at 05:00 UTC
- Released
- 31 Mar 2026
- Status
- Current
- Context
- 262.1K
- Max output
- —
- Input per 1M tokens
- —
- Output per 1M tokens
- —
- Open weights
- Yes
- License
- apache-2.0
- Parameters
- 25.2B, 3.8B active
- Knowledge cutoff
- Jan 2025
- Reasoning
- Yes
- Modalities
- text, image, video → text
Category scores
Out of 100, from independent and third-party results only, by method category-2026-09-29.1.
| Category | Score | Rank | Coverage |
|---|---|---|---|
| Knowledge & reasoning | 18.0 | 82 | 2 of 4 |
| Human preference | 42.7 | 40 | 3 of 5 |
Benchmark results
Every result we hold for this model, with who ran it and where it was published.
Knowledge & reasoning
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| Chess Puzzles | Puzzles solved | 6.0% | Independent | minimal | 6 Aug 2026 | Epoch AI |
| DTBench (Decision Theory Benchmark) | Accuracy (capability questions) | 74.9% | Third party | — | — | https://conceptualreasoning.ai/dtbench via Epoch AI |
| LMCA (Language Model Conceptual Argumentation) | Correlation with expert ratings (Pearson r) | 0.2968 | Third party | — | — | https://conceptualreasoning.ai/lmca via Epoch AI |
Scientific reasoning
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| GPQA Diamond | Accuracy | 73.2% | Independent | minimal | 6 Aug 2026 | Epoch AI |
Mathematics
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| OTIS Mock AIME 2024-2025 | Accuracy | 82.2% | Independent | minimal | 28 Aug 2026 | Epoch AI |
Coding & software engineering
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| WeirdML (v2) | Average accuracy | 35.2% | Third party | — | — | https://htihle.github.io/weirdml.html via Epoch AI |
Human preference
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| Code Arena: WebDev (arena.ai) | Arena score (Elo scale) | 1358 | Independent | — | 25 Sep 2026 | Arena |
| Code Arena: WebDev (arena.ai) | Votes | 1505 | Independent | — | 25 Sep 2026 | Arena |
| Text Arena (arena.ai) | Arena score (Elo scale) | 1438 | Independent | — | 25 Sep 2026 | Arena |
| Text Arena (arena.ai) | Votes | 6078 | Independent | — | 25 Sep 2026 | Arena |
| Vision Arena (arena.ai) | Arena score (Elo scale) | 1242 | Independent | — | 28 Sep 2026 | Arena |
| Vision Arena (arena.ai) | Votes | 23977 | Independent | — | 28 Sep 2026 | Arena |
Other Gemma versions
- Gemma 4 12B Unified3 Jun 2026
- Gemma 4 31B31 Mar 2026
- Gemma 4 E2B31 Mar 2026
- Gemma 4 E4B31 Mar 2026
- Gemma 3 270M14 Aug 2025
- Gemma 3n E2B26 Jun 2025
- Gemma 3n E4B26 Jun 2025
- Gemma 3 12B10 Mar 2025
- Gemma 3 1B10 Mar 2025
- Gemma 3 27B10 Mar 2025
- Gemma 3 4B10 Mar 2025
- Gemma 2 JPN 2B3 Oct 2024
- Gemma 2 2B31 Jul 2024
- Gemma 2 27B27 Jun 2024
- Gemma 2 9B27 Jun 2024
- Gemma 1.1 2B5 Apr 2024
- Gemma 1.1 7B5 Apr 2024
- Gemma 2B21 Feb 2024
- Gemma 7B21 Feb 2024