Cerebras·released 2023-03-201 source
Cerebras-GPT-13B benchmark scores: 6 benchmarks tracked. 0% of its scores are independently reproduced — the rest are self-reported or unverified.
Leads on 0 of 6 benchmarks·best result 47 on PIQA (self-reported)·0 of 6 independently reproduced
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| PIQA | 47accuracy (%) | self-reported· optimizedT1 | 2023-03-20XGen-7B Technical Report |
| HellaSwag | 45.87accuracy (%) | self-reported· optimizedT1 | 2023-03-20XGen-7B Technical Report |
| Winogrande | 21.6accuracy (%) | self-reported· optimizedT1 | 2023-03-20XGen-7B Technical Report |
| OpenBookQA | 14.4accuracy (%) | self-reported· optimizedT1 | 2023-03-20XGen-7B Technical Report |
| ARC AI2 | 9.87accuracy (%) | self-reported· optimizedT1 | 2023-03-20XGen-7B Technical Report |
| MMLU | 1.6accuracy (%) | self-reported· optimizedT1 | 2023-03-20XGen-7B Technical Report |
Claims drawn from cited facts, not live model generation.
Developed by Cerebras and released in 2023, this model is a mid-size 13-billion-parameter system, a scale that balances meaningful capability with lower operational cost than frontier-scale alternatives.
3 cited facts
This model is scored on 6 tracked benchmarks, but currently holds no top score on any of them. Because the record is self-reported, these numbers are vendor-claimed and not yet independently confirmed, so they should be read as claims rather than verified results.
3 cited facts
With an average SOTA gap of 62.12 points, the model trails the leader by a wide margin and sits well behind the front of the pack on its benchmarks. It holds the top score on none of the tracked SOTA benchmarks (leader count 0), so it has 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
Cerebras-GPT-13B is an AI model developed by Cerebras, released 2023-03-20. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.
Cerebras-GPT-13B has recorded scores on 6 benchmarks, each shown with its evidence status.
Cerebras-GPT-13B has recorded scores on 6 benchmarks — PIQA, OpenBookQA, Winogrande, MMLU, ARC AI2, HellaSwag. The full table above shows each score with its evidence status.
0 of 6 recorded scores (0%) are independently reproduced rather than self-reported by the lab.
Cerebras-GPT-13B has 6 tracked claims: 6 self-reported.