Meta AI·released 2023-02-241 source
LLaMA-33B 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 83.8 on TriviaQA (unverified)·0 of 10 independently reproduced
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| TriviaQA | 83.8accuracy (%) | unverified· optimizedT1 | 2023-02-24Llama 2: Open Foundation and Fine-Tuned Chat Models |
| LAMBADA | 77.2accuracy (%) | self-reported· optimizedT1 | 2023-02-24Qwen Technical Report |
| HellaSwag | 77.07accuracy (%) | unverified· optimizedT1 | 2023-02-24LLaMA: Open and Efficient Foundation Language Models |
| PIQA | 64.6accuracy (%) | self-reported· optimizedT1 | 2023-02-24Qwen Technical Report |
| ARC AI2 | 56.67accuracy (%) | self-reported· optimizedT1 | 2023-02-24Qwen Technical Report |
| Winogrande | 52accuracy (%) | unverified· optimizedT1 | 2023-02-24LLaMA: Open and Efficient Foundation Language Models |
| MMLU | 44.93accuracy (%) | unverified· optimizedT1 | 2023-02-24Llama 2: Open Foundation and Fine-Tuned Chat Models |
| OpenBookQA | 44.8accuracy (%) | unverified· optimizedT1 | 2023-02-24Llama 2: Open Foundation and Fine-Tuned Chat Models |
| GSM8K | 44.1accuracy (%) | unverified· optimizedT1 | 2023-02-24Stanford HELM |
| BBH | 33.33accuracy (%) | self-reported· optimizedT1 | 2023-02-24Qwen Technical Report |
Claims drawn from cited facts, not live model generation.
Developed by Meta AI, this model was released in 2023, establishing its origin within Meta's AI research efforts. With 33 billion parameters, it is a large-scale model, offering substantial headroom for complex tasks while also demanding greater computational resources and cost.
3 cited facts
This model is tracked across 10 benchmarks and currently holds no top score in any of them. These figures are self-reported and have not been independently confirmed, so treat them as claims rather than verified results.
3 cited facts
With an average gap of 27.96 points behind the SOTA score, this model sits well behind the leaders on the benchmarks measured. It holds the top score on none of the benchmarks (0), consistent with a wide-gap, non-frontier position.
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
LLaMA-33B is an AI model developed by Meta AI, released 2023-02-24. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.
LLaMA-33B has recorded scores on 10 benchmarks, each shown with its evidence status.
LLaMA-33B 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.
LLaMA-33B has 10 tracked claims: 4 self-reported, 6 unverified.