Benchmark
NoCha pair accuracy (ordinary-model-common)
Long context benchmarks measure how well models use information in long inputs. Scores are percentages. Higher is better.
- Models and systems measured
- 6
- Scores recorded
- 6
- Sources
- 1
- IQ weight
- No weight published
Score distribution
- 45.5%
- Mean
- 48.3%
- Median
- 7.7%
- Std dev
- 40.1%
- 10th percentile
- 58.2%
- 90th percentile
- 35%
- Min
Results
The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.
| Model or system | Score |
|---|---|
| Claude-3-Opustable: 3, view: ordinary-model-common, system: Claude-3-Opus· NoCha (arXiv:2406.16264v3) data | 50.8% |
| Claude-3.5-Sonnettable: 3, view: ordinary-model-common, system: Claude-3.5-Sonnet· NoCha (arXiv:2406.16264v3) data | 40.7% |
| Gemini Flash 1.5table: 3, view: ordinary-model-common, system: Gemini Flash 1.5· NoCha (arXiv:2406.16264v3) data | 35% |
| Gemini Pro 1.5table: 3, view: ordinary-model-common, system: Gemini Pro 1.5· NoCha (arXiv:2406.16264v3) data | 48.3% |
| GPT-4-Turbotable: 3, view: ordinary-model-common, system: GPT-4-Turbo· NoCha (arXiv:2406.16264v3) data | 40.1% |
| GPT-4otable: 3, view: ordinary-model-common, system: GPT-4o· NoCha (arXiv:2406.16264v3) data | 58.2% |
Data credits: Marzena Karpinska, Katherine Thai, Kyle Lo, Tanya Goyal and Mohit Iyyer. One Thousand and One Pairs, arXiv:2406.16264v3. CC BY 4.0 article numeric/protocol extract; restructured, printed values unchanged. No task, book or dataset rights asserted. (licence)