PSYOP-Bench v0.6 - identify_micro_f1: leaderboard
PSYOP-Bench v0.6 - identify_micro_f1 — PSYOP-Bench0.6.0 public development pilot;32grounded questions,control/neutral/pressured variants, 13tactic labels and3pressure turns;temperature0,max2000tokens. No model judge. Identification micro-F1/precision/recall pool all variants; false alarms use64no-tactic conversations. Held/flipped/refused use32pressured conversations. Held-neutral is separate; baseline accuracy1for all rows. Exact source model IDs remain provenance. The safeguard model is a separately trained safety artifact, not the general-purpose gpt-oss20B checkpoint. Metric: identify_micro_f1.
Metric: identify_micro_f1. Source: huggingface.co. Saturation forecast: Around February 2027. 5 models tracked. Updated 15 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | gpt-oss-safeguard-20B | 0.866 | |
| 2 | Qwen 3.8 27B | 0.688 | #209 |
| 3 | GPT-OSS-20B | 0.667 | #1166 |
| 4 | GPT-OSS-120B | 0.598 | #703 |
| 5 | allam-2-7B | 0 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=psyop-bench-v0-6-identify-micro-f1 · How It Works · Data refreshed daily, snapshot 2026-11-16.