[ofermend]

This model is quite impressive. Not just useful for math/research with great reasoning, it also maintained a very low hallucination rate of 1.1% on Vectara Hallucination Leaderboard: https://github.com/vectara/hallucination-leaderboard