Model card
Gemma 3 12B IT.
Google DeepMindopen-weights12B paramstransformer
Gemma 3 instruction-tuned, March 2025
§ 01 · Benchmarks
Every benchmark Gemma 3 12B IT has a recorded score for.
| # | Benchmark | Area · Task | Metric | Value | Rank | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | MBPP | Computer Code · Code Generation | pass@1 | 73.0% | #15 | — | source ↗ |
| 02 | LiveCodeBench | Computer Code · Code Generation | pass@1 | 32.0% | #28 | — | source ↗ |
| 03 | HumanEval | Computer Code · Code Generation | pass@1 | 85.4% | #29 | — | source ↗ |
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 02 · Strengths by area
Where Gemma 3 12B IT actually performs.
§ 04 · Related models
Other Google DeepMind models scored on Codesota.
§ 05 · Sources & freshness
Where these numbers come from.
arxiv-2503.19786
3
results
3 of 3 rows marked verified.