Here we go again: a new stealth model scores 74% on DeepSWE, beating GPT-5.6 Sol and most other models.
It also reportedly outperforms GPT-5.6 Sol on Terminal-Bench 2.1 and SWE-Bench Verified, and being much cheaper at the same time. My bet is on another chinese model.
Here's the DeepSWE result for Union Alpha, a stealth model we just launched. Try it now! Works in every harness. $ ori [code | your-fav-harness] --model=stealth...