Google DeepMind·released 2024-02-211 source
Gemma 7B benchmark scores: 10 benchmarks tracked. 0% of its scores are independently reproduced — the rest are self-reported or unverified.
Leads on 0 of 10 benchmarks·best result 76.27 on HellaSwag (unverified)·0 of 10 independently reproduced
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| HellaSwag | 76.27accuracy (%) | unverified· optimizedT1 | 2024-02-21Gemma: Open Models Based on Gemini Research and Technology |
| TriviaQA | 72.3accuracy (%) | unverified· optimizedT1 | 2024-02-21Gemma: Open Models Based on Gemini Research and Technology |
| OpenBookQA | 71.47accuracy (%) | self-reported· optimizedT1 | 2024-02-21Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone |
| ARC AI2 | 71.07accuracy (%) | self-reported· optimizedT1 | 2024-02-21Nemotron-4 15B Technical Report |
| PIQA | 62.4accuracy (%) | unverified· optimizedT1 | 2024-02-21Gemma: Open Models Based on Gemini Research and Technology |
| Winogrande | 58accuracy (%) | unverified· optimizedT1 | 2024-02-21Gemma: Open Models Based on Gemini Research and Technology |
| MMLU | 54.8accuracy (%) | unverified· optimizedT1 | 2024-02-21Stanford CRFM Leaderboard |
| GSM8K | 46.4accuracy (%) | unverified· optimizedT1 | 2024-02-21Gemma: Open Models Based on Gemini Research and Technology |
| BBH | 40.13accuracy (%) | unverified· optimizedT1 | 2024-02-21Gemma: Open Models Based on Gemini Research and Technology |
| ANLI | 23.05accuracy (%) | self-reported· optimizedT1 | 2024-02-21Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone |
Claims drawn from cited facts, not live model generation.
Developed by Google DeepMind, this model is a compact 7-billion-parameter system released in 2024, which implies greater efficiency and lower cost rather than maximum headroom.
3 cited facts
This model is scored on 10 tracked benchmarks, and it currently holds no top score on any of them. The benchmark record is self-reported, meaning the numbers are vendor-claimed and not yet independently confirmed, so they should be read as claims rather than verified results.
3 cited facts
It trails the state-of-the-art by an average of 23.96 points, a wide margin that reads as well behind the leaders. It holds the top score on 0 benchmarks, indicating no category leadership yet.
2 cited facts
2 facts cross-checked across data sources: 2 single-source.
Source facts, citations, and refresh stamp for this record.
Sources: curated_param_counts, epoch_benchmarks
Gemma 7B is an AI model developed by Google DeepMind, released 2024-02-21. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.
Gemma 7B has recorded scores on 10 benchmarks, each shown with its evidence status.
Gemma 7B has recorded scores on 10 benchmarks — PIQA, OpenBookQA, Winogrande, TriviaQA, MMLU, ARC AI2, and 4 more. The full table above shows each score with its evidence status.
0 of 10 recorded scores (0%) are independently reproduced rather than self-reported by the lab.
Gemma 7B has 10 tracked claims: 3 self-reported, 7 unverified.