BLIP-2-OPT-2.7B: benchmark results
Provider: Other.
Aggregate Rating 1400 ± 52, rank #6172 of 10489 rated models, from 23 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CoreCognition Image only Conservation | 61.27 | Accuracy (%) | 78.3 |
| Epoch AI - Science Qa | 74.17 | Score | 78.3 |
| CoreCognition Image only Perceptualconstancy | 57.81 | Accuracy (%) | 50 |
| CoreCognition Image only Perspective | 32.82 | Accuracy (%) | 38.2 |
| Mind's Eye - Hierarchical Pattern Equivalence | 22.7 | Accuracy (%) on the hierarchical pattern equivalence task of | 35.3 |
| CoreCognition Image only Spatiality | 41.48 | Accuracy (%) | 17.8 |
| Mind's Eye - Symmetric Structures | 17 | Accuracy (%) on the symmetric structures task of Mind's Eye | 17.6 |
| CoreCognition Image only Permanence | 21.88 | Accuracy (%) | 17.1 |
| CoreCognition Image only Hierarchy | 30.83 | Accuracy (%) | 11.8 |
| Mind's Eye - Dynamic Structural Correspondence | 18.3 | Accuracy (%) on the dynamic structural correspondence task o | 11.8 |
| Mind's Eye - Mental Composition | 24 | Accuracy (%) on the mental composition (net to solid) task o | 11.8 |
| CoreCognition Image only Boundary | 45.41 | Accuracy (%) | 10.5 |
Interactive version: theaggregate.ai/model?slug=blip-2-opt-2-7b · How It Works · Data refreshed daily, snapshot 2026-11-20.