Updated daily at 05:00 UTC
Benchmark
ScienceQA
Measures: Multiple-choice science questions from grade 1 to 12 curricula, many with an image or diagram, annotated with lectures and explanations.
- Maintainer
- Lu et al. (UCLA and AI2)
- Introduced
- 2022
- Status
- Unmaintained
- Saturation
- Saturated
- Typical provenance
- Mixed
- In scoring set
- No
Leaderboard
Independent and third-party results only.
| Model | Accuracy | Provenance | Source | Evaluated |
|---|---|---|---|---|
| text-davinci-001 (InstructGPT) | 74.0% | Third party | ScienceQA leaderboard via Epoch AI | — |
| text-davinci-001 (InstructGPT) | 74.0% | Third party | ScienceQA leaderboard via Epoch AI | — |
| Qwen-VL-Chat | 68.2% | Third party | VisionLLM v2: An End-to-End Generalist Multimodal Large Language Model for Hundreds of Vision-Language Tasks via Epoch AI | — |
| Llama 2 13B | 55.8% | Third party | ScienceQA leaderboard via Epoch AI | — |
| Llama 2 7B | 43.1% | Third party | ScienceQA leaderboard via Epoch AI | — |
- Epoch AI, 'Capabilities & benchmarking', epoch.ai/benchmarks, CC BY 4.0.