AI BriefingAIM System One Index ranks 57 decision models on accuracy, bias, cost, and latency
·2026.10.21 08:09
Key point
AutoJev-27B, d1-ultra, and decider-12b tie for the top spot with a score of 60 in the new benchmark for single-pass decision models.
Details
- AIM System One Index is a new leaderboard ranking 57 "System One" models that return typed decisions with probabilities in a single pass rather than generating text token by token.
- Three models tie for the top score of 60: AutoJev-27B, d1-ultra, and decider-12b.
- d1-ultra achieves the best browser task score at 74% and costs $0.026 per 1,000 decisions.
- Laya is the fastest model at 30 ms per decision, while Decider Flash V1.2 is the cheapest at $0.007 per 1,000 decisions.
- The index combines three benchmarks: 50 browser tasks, 1,655 single-decision classification questions, and bias tests with rotated "cannot be determined" options.
- Scores are z-scores where 50 is average and 10 points equals one standard deviation; cost and latency are displayed but excluded from the index score.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Mercor Benchmark Map Reveals 90% of US Working Hours Lack Adequate AI Benchmark CoverageCheck Point Finds Jev Decision Model Vulnerable to Prompt Injection for ~$0.50
AI Benchmarks: Climbing the Wrong Hills