GSM8K#3 of 28
exact_match
View run →95.6%±0.6%
Evaluated across 13 benchmarks. Ranks in the top 3 on 5 of 13. Strongest showing on GSM8K (95.6% exact_match, #3 of 28). Weakest on MBPP(Instruct) (0.0% pass@1, #14 of 24).
| Benchmark | Metric | Score | Rank | Actions |
|---|---|---|---|---|
| GSM8K | exact_match | 95.6%±0.6% | #3 of 28 | View run → |
| IFEval | prompt_level_strict_acc | 89.5%±1.3% | #4 of 28 | View run → |
| EQ-Bench | eqbench | 84.2±1.2 | #4 of 28 | View run → |
| MMLU-Pro | exact_match | 82.3%±0.3% | #3 of 28 | View run → |
| MGSM | exact_match | 74.5%±0.7% | #6 of 28 | View run → |
| BBH | exact_match | 48.1%±0.4% | #9 of 28 | View run → |
| MMLU | acc | 47.9%±0.4% | #22 of 28 | View run → |
| GPQA Extended | acc | 44.3%±2.1% | #3 of 28 | View run → |
| GPQA Main | acc | 41.7%±2.3% | #4 of 28 | View run → |
| GPQA Diamond | acc | 41.4%±3.5% | #3 of 28 | View run → |
| LongBench | score | 40.2%±0.5% | #3 of 10 | View run → |
| MBPP | pass@1 | 24.4%±1.9% | #17 of 28 | View run → |
| MBPP(Instruct) | pass@1 | 0.0%±0.0% | #14 of 24 | View run → |
Citation
FrozeBench. "google/gemma-4-26B-A4B-it." https://frozebench.com/models/google-gemma-4-26b-a4b-it. Retrieved 2026-06-04.
BibTeX
@misc{frozebench_google_gemma_4_26B_A4B_it,
title = {google/gemma-4-26B-A4B-it},
howpublished = {\url{https://frozebench.com/models/google-gemma-4-26b-a4b-it}},
year = {2026},
note = {FrozeBench. Retrieved 2026-06-04.}
}URL
https://frozebench.com/models/google-gemma-4-26b-a4b-it