Zhu Liang(@paradite_):this tb4 story keeps getting wilder. both vals ai and aa use mini-swe-agent harness, but the ranking is completely different: vals ai: opus 5.5 > sonnet 5.5 > kimi k3.1 pro > gemini 4 carbon aa: kimi k3.1 pro > gemini 4 carbon> sonnet 5.5

Zhu Liang

@paradite_

building now: built in the past: @ai_simulator @16xPrompt @16xEval

参加 August 2009

278 フォロー中 5.2K ファン

Zhu Liang@paradite_

2026.10.29 17:43

this tb4 story keeps getting wilder. both vals ai and aa use mini-swe-agent harness, but the ranking is completely different: vals ai: opus 5.5 > sonnet 5.5 > kimi k3.1 pro > gemini 4 carbon aa: kimi k3.1 pro > gemini 4 carbon> sonnet 5.5 > opus 5.5 you can't make this up.