Update README.md

README.md

@@ -40,7 +40,11 @@ configs:
 - **Leaderboard** [https://grab-benchmark.github.io](https://grab-benchmark.github.io)
 
 ### Dataset Summary
-TODO
+Large multimodal models (LMMs) have exhibited proficiences across many visual tasks. Although numerous benchmarks exist to evaluate model performance, they increasing have insufficient headroom and are **unfit to evaluate the next generation of frontier LMMs**.
+
+To overcome this, we present **GRAB**, a challenging benchmark focused on the tasks **human analysts** might typically perform when interpreting figures. Such tasks include estimating the mean, intercepts or correlations of functions and data series and performing transforms.
+
+We evaluate a suite of **20 LMMs** on GRAB, finding it to be a challenging benchmark, with the current best model scoring just **21.7%**.
 
 ### Example usage
 ```python