MATH-500
Harbor adapterMATH-500: 500道精选竞赛数学题
Verified100%Math
- Samples
- 500
- Released
- 2023
- Scores
- 9
MATH-V
Math-Vision (Math-V) dataset is a meticulously curated collection of 3,040 high-quality mathematical problems with visual contexts sourced from real math competitions.
How to run
ltb · math-v
$ npm i -g lettokenburn-cli$ ltb bench info math-v
External harness. No Harbor adapter has been generated for this benchmark yet. `ltb bench info` prints the known links.
Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs
Scores
4 sourced entries| Harness | Source | Confidence | Link | |||
|---|---|---|---|---|---|---|
| GPT-4V | 22.8 | N/A | Independent | medium | Feb 22, 2024 | arxiv.org |
| Gemini Pro | 17.7 | N/A | Independent | medium | Feb 22, 2024 | arxiv.org |
| Qwen-VL-Max | 15.6 | N/A | Independent | medium | Feb 22, 2024 | arxiv.org |
| InternLM-XComposer2-VL | 14.5 | N/A | Independent | medium | Feb 22, 2024 | arxiv.org |
Every number links to where it was published. Low-confidence rows are muted and need per-cell verification.
Reproduced by the community
All runsLoading community runs…
About
Math-Vision (Math-V) dataset is a meticulously curated collection of 3,040 high-quality mathematical problems with visual contexts sourced from real math competitions. Spanning 16 distinct mathematical disciplines and graded across 5 levels of difficulty, our dataset provides a comprehensive and diverse set of challenges for evaluating the mathematical reasoning abilities of LMMs.
Through extensive experimentation, we unveil a notable performance gap between current LMMs and human performance on Math-Vision, underscoring the imperative for further advancements in LMMs. Moreover, our detailed categorization allows for a thorough error analysis of LMMs, offering valuable insights to guide future research and development.