When looking at the benchmarks, this model seems to be really close to Kimi K2.6 in terms of intelligence and pricing, hitting that sweet spot. It does also have a higher AA-Omniscience index, which is something kimi and other open models lack in. Curious to see how pleasant it is to use.