CIPHER Cross-Record Inference - FIR (Program-of-Thought): leaderboard

Metric: Harmonic answer score (%; per question the harmonic mean of token overlap with the reference and Gemini 2.5 Flash semantic acceptance, zero when rejected, averaged; First Information Reports (Indian police records) with typed placeholders from post-hoc masking; the full support-controlled context of 39-42 records including distractors; program-of-thought prompting; greedy decoding). Source: arxiv.org. Saturation forecast: Around January 2028. 5 models tracked.

Top models

#ModelScore
1Qwen 3 Next 80B A3B33.9
2GPT-OSS-20B29.2
3Gemini 2.0 Flash13.4
4Llama 3.3 70B Instruct10.1
5GPT-OSS-120B5.9

Interactive version: theaggregate.ai/benchmark?slug=cipher-cross-record-inference-fir-program-of-thought · How It Works · Data refreshed daily, snapshot 2026-09-26.