GSM8K#2 of 28
exact_match
View run →96.6%±0.5%
Benchmark results loaded.
google/gemma-4-31B-it
- Weights
- 62.6 GB · safetensors
- Source
- huggingface page
- Commit
- fb9ae26
Evaluated across 13 benchmarks. Ranks in the top 3 on 10 of 13. Strongest showing on GSM8K (96.6% exact_match, #2 of 28). Weakest scores tied at 0.0% on MBPP, and MBPP(Instruct).
Benchmark results
| Benchmark | Metric | Score | Rank | Actions |
|---|---|---|---|---|
| GSM8K | exact_match | 96.6%±0.5% | #2 of 28 | View run → |
| IFEval | prompt_level_strict_acc | 91.3%±1.2% | #1 of 28 | View run → |
| EQ-Bench | eqbench | 85.8±1.2 | #2 of 28 | View run → |
| MMLU-Pro | exact_match | 84.9%±0.3% | #1 of 28 | View run → |
| MMLU | acc | 82.7%±0.3% | #3 of 28 | View run → |
| BBH | exact_match | 78.9%±0.3% | #2 of 28 | View run → |
| MGSM | exact_match | 76.6%±0.7% | #3 of 28 | View run → |
| GPQA Diamond | acc | 52.0%±3.6% | #2 of 28 | View run → |
| GPQA Extended | acc | 51.6%±2.1% | #1 of 28 | View run → |
| GPQA Main | acc | 50.4%±2.4% | #2 of 28 | View run → |
| LongBench | score | 37.0%±0.4% | #4 of 10 | View run → |
| MBPP | pass@1 | 0.0%±0.0% | #27 of 28 | View run → |
| MBPP(Instruct) | pass@1 | 0.0%±0.0% | #15 of 24 | View run → |
- IFEval#1 of 28prompt_level_strict_accView run →91.3%±1.2%
- EQ-Bench#2 of 28eqbenchView run →85.8±1.2
- MMLU-Pro#1 of 28exact_matchView run →84.9%±0.3%
- MMLU#3 of 28accView run →82.7%±0.3%
- BBH#2 of 28exact_matchView run →78.9%±0.3%
- MGSM#3 of 28exact_matchView run →76.6%±0.7%
- GPQA Diamond#2 of 28accView run →52.0%±3.6%
- GPQA Extended#1 of 28accView run →51.6%±2.1%
- GPQA Main#2 of 28accView run →50.4%±2.4%
- LongBench#4 of 10scoreView run →37.0%±0.4%
- MBPP#27 of 28pass@1View run →0.0%±0.0%
- MBPP(Instruct)#15 of 24pass@1View run →0.0%±0.0%
last evaluated:3 weeks ago
How to cite
Citation
FrozeBench. "google/gemma-4-31B-it." https://frozebench.com/models/google-gemma-4-31b-it. Retrieved 2026-06-04.
BibTeX
@misc{frozebench_google_gemma_4_31B_it,
title = {google/gemma-4-31B-it},
howpublished = {\url{https://frozebench.com/models/google-gemma-4-31b-it}},
year = {2026},
note = {FrozeBench. Retrieved 2026-06-04.}
}URL
https://frozebench.com/models/google-gemma-4-31b-it