Grok 4.5 about a month ago was disappointing, not in an absolutely bad way but rather in a vague position feeling. This model seems to have been improved through post-training without increasing the number of parameters. It has improved significantly in benchmarks. The price is also much cheaper compared to Claude.
As I recall from previous articles, there was mention that models like recent Chinese super large language models were also being prepared, so I'm looking forward to the next version. I thought it would be better to go down and be overtaken by Chinese models after being in 3rd or 4th place for years
▶ Original source: https://x.ai/news/grok-4-6
