Google
availableShows if the model has enough results for an index.Gemma 4 12B
Gemma 4 12B is a reasoning model from Google in the Gemma 4 family. 18 benchmarks count toward its score, in 8 categories.
IndexOverall score out of 100.41.7 ±3.8
CoverageShare of the index weight with results.100%
SpeedOutput tokens per second.—
Input / 1MUS dollars per 1M input tokens.N/A
Output / 1MUS dollars per 1M output tokens.N/A
ContextMaximum tokens in one request.256K
EloLMArena rating and rank.N/A
The index is a score out of 100. The ± range shows how much it can change.
CapabilitiesScore per category, out of 100.
Out of 100Results
18 counted| BenchmarkThe test name. | CategoryThe capability that the test measures. | ResultThe score from the publisher. | IndexThis result as a score out of 100. | RunThe settings of the run. | DateDate of the result. | Published byThe source of the result. |
|---|---|---|---|---|---|---|
| MMMLU | Knowledge | 83.4% | — | — | — | OpenAI |
| MathVision | Multimodal | 79.7% | — | — | — | Qwen |
| Graduate-Level Google-Proof Q&A | Knowledge | 78.8% | 51.0 | — | — | David Rein et al. |
| GPQA Diamond | Knowledge | 78.8% | 51.0 | — | — | David Rein et al. |
| AIME 2026 | Math | 77.5% | 41.1 | — | — | Qwen |
| Massive Multitask Language Understanding Professional | Knowledge | 77.2% | 42.1 | — | — | Yubo Wang et al. |
| Artificial Analysis GPQA Diamond | Knowledge | 75.3% | 45.7 | — | — | Artificial Analysis |
| Artificial Analysis IFBench | Instruction | 73.5% | 63.7 | — | — | Artificial Analysis |
| LiveCodeBench v6 | Coding | 72.0% | 42.6 | — | — | LiveCodeBench maintainers |
| Artificial Analysis MMMU-Pro | Multimodal | 69.7% | 52.3 | — | — | Artificial Analysis |
| Massive Multi-discipline Multimodal Understanding Pro | Multimodal | 69.1% | 39.8 | — | — | MMMU-Pro authors |
| Artificial Analysis Long Context Reasoning | Reasoning | 63.7% | 52.3 | — | — | Artificial Analysis |
| BIG-Bench Hard | Reasoning | 53.0% | — | — | — | Mirac Suzgun et al. |
| MedXpertQA Multimodal | Multimodal | 48.7% | — | — | — | Meta AI |
| EuroEval Portuguese | Multilingual | 45.5% | 59.1 | — | — | EuroEval |
| EuroEval Dutch | Multilingual | 44.9% | 58.3 | — | — | EuroEval |
| MRCRv2 | Reasoning | 43.4% | — | — | — | OpenAI |
| EuroEval Polish | Multilingual | 43.1% | 56.1 | — | — | EuroEval |
| EuroEval French | Multilingual | 42.8% | 55.7 | — | — | EuroEval |
| EuroEval Swedish | Multilingual | 42.0% | 54.8 | — | — | EuroEval |
| EuroEval Italian | Multilingual | 41.5% | 54.1 | — | — | EuroEval |
| EuroEval German | Multilingual | 37.3% | 48.9 | — | — | EuroEval |
| τ²-Bench Tool-Agent-User Evaluation | Agentic | 36.3% | 24.9 | — | — | Victor Barres et al. |
| EuroEval Spanish | Multilingual | 32.3% | 42.7 | — | — | EuroEval |
| Artificial Analysis Coding Index | Coding | 31.0% | 40.8 | — | — | Artificial Analysis |
| Artificial Analysis Humanity's Last Exam | Knowledge | 15.7% | 41.6 | — | — | Artificial Analysis |
| Artificial Analysis Omniscience Accuracy | Knowledge | 15.6% | 33.2 | — | — | Artificial Analysis |
| Artificial Analysis Intelligence Index | Knowledge | 14.2% | 40.1 | — | — | Artificial Analysis |
| Humanity's Last Exam without tools | Knowledge | 5.2% | 33.2 | — | — | OpenAI |
| GDPval-AA normalized | Agentic | 0.0% | 35.6 | — | — | Artificial Analysis |
| Critical Physics Tasks | Reasoning | 0.0% | 38.4 | — | — | Artificial Analysis |
18 benchmarks count, from 26 of 31 results. A grey row does not count. Too few models took that benchmark.