added notes on benchmark

README.md

@@ -43,11 +43,20 @@ It has been trained specifically for these tasks:
 * natural product elucidation (formula + organism to SMILES)
 * blood-brain barrier permeability
 
-<img src="./images/benchmarks.png" width="800">
-
 For example, you can ask "Propose a molecule with a pKa of 9.2" or "Modify CCCCC(O)=OH to increase its pKa by about 1 unit." You cannot ask it "What is the pKa of CCCCC(O)=OH?"
 If you ask it questions that lie significantly beyond those tasks, it can fail. You can combine properties, although we haven't significantly benchmarked this.
 
+## Benchmarks
+
+We tested ether0, along with some experts and frontier models, on [a benchmark we developed](https://huggingface.co/datasets/futurehouse/ether0-benchmark/).
+The benchmark is made from commonly used tasks - like reaction prediction in USPTO, molecular captioning from PubChem, or predicting GHS classification.
+The benchmark is different in two ways: all answers are a molecule, and we balanced it so that each task is 25 questions so that is reasonable length for frontier model evals.
+The tasks generally follow previously reported numbers - e.g., a reaction prediction accuracy of 80% here would be about the same on a withheld split of the USPTO-50k dataset.
+
+The results below are the model weights released in this repo. This is different than the preprint, which has pre-safety mitigation benchmarks.
+
+<img src="./images/benchmarks.png" width="800">
+
 ## Limitations
 
 It does not know general synonyms and it has poor textbook knowledge (e.g. it does not perform especially well on chembench).