The benchmark confirms that some AIs are becoming exceptional tools for research-level math. But open-source models still struggle
The benchmark confirms that some AIs are becoming exceptional tools for research-level math. But open-source models still struggle