This is a preview model by Tencent, announced on August 28, 2026.
I have not run the full LLM coding benchmark with it, because, after just 1 prompt out of 20, the time/cost appeared to be much higher than its predecessor Hy3.
In comparison with this $0.41 cost for one prompt that took 22 minutes to run, Tencent Hy3 numbers for the same prompt were ~$0.05 and ~6 minutes, averaged from 5 attempts.
So, for now, I decided to NOT proceed with full benchmark, as it seems that it would be too expensive, so I can't recommend it to my audience.
