{
  "id": 14106346,
  "title": "Decision AI Global Benchmark S1MB No. 1 : Perplexity · Cloudflare · Meta beaten by Darwin-27B-ZTC-v5",
  "url": "https://urgent.news/2026/10/13/ai-s1mb-1-darwin-27b-ztc-v5",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-13T03:28:39.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/vidraft_lab/disijeon-ai-geulrobeol-bencimakeu-s1mb-1wi-peopeulregsitikeulraudeupeulreeometa-jecin-darwin-27b-ztc-v5-49l8"
  },
  "original_language": "ko",
  "account": "비드래프트의 판정 엔진 Darwin-27B-ZTC-v5가 허깅페이스 공개 벤치마크 S1MB에서 152개 모델 중 종합 1위에 올랐다. 이 모델은 Borda 점수 91.16을 기록하며 2위인 퍼플렉시티의 pplx-decider-v1.1-27b를 2.50점 앞섰다. Darwin-27B-ZTC-v5는 특히 점수 매기기 부문에서 큰 차이로 경쟁 모델들을 앞섰으며, 타입세이프AI의 공식 모델 Jev 1.13.0과는 9.72점 차이로 1위에 올랐다. 이 모델은 현재 허깅페이스 트랜스포머스의 판정 서비스 기능에서 별도 코드 없이 실행 가능하다.",
  "summary": "TL;DR: The non-draft judgment engine, Darwin-27B-ZTC-v5, ranked first overall among 152 models in the Hugging Face public benchmark S1MB. With a Borda score of 91.16, calculated by converting 137 benchmark rankings into scores and then averaging them, it outperformed Perplexity (88.66) in second place by 2.50 points and TypeSafe AI's JEV official model (ranked 14th) by 9.72 points. Two Beedraft models secured positions within the top 5, with one in first place and the other in fourth. What is S1MB and why is it important? S1MB (System One Mosaic Benchmark) is a public leaderboard for evaluating Decision AI. Decision AI refers to AI that does not generate text but makes judgments. Given a situation and a question, it provides probabilities for each option as numerical values, allowing programs to directly incorporate these values into conditional statements.",
  "key_points": [],
  "editors_take": "Bidraft’s dominance in the S1MB benchmark signals a shift toward specialized decision-making models that enable software to execute immediate, systemic logic through precise numerical probability outputs.",
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}