Verification Transport - PlanRewardBench (PRB) - V-LLM - Delegation Restoration Recovery: leaderboard

Matched native/abstracted/restored executions; 7exactmodelbackbones; protocol andbenchmark split. OBA nativebalancedaccuracy; VTFR conditionedonnativecorrectjudgments; RRR conditionedontransportfailures; RDR restoreddegradation. Denominators differ andmustnotpool. Invalid/noverdict outcomes countedunderpaperoperationalfailurepolicy. Three abstractionboundaries independent. Full immutablepayload pins alldefinitions. Metric: Native rate (%). Source: https://arxiv.org/html/2611.18471v1

Metric: Native rate (%). Source: arxiv.org. Saturation forecast: Not forecast. 7 models tracked. Updated 17 Nov 2026.

Top models

#ModelScoreOverall rank
1Llama 4 Maverick94.7#1118
2Gemma 4 31B92#385
3GPT-OSS-120B86.2#707
4Gemma 4 26B A4B85.1#581
5GPT-OSS-20B81.5#1163
6Gemini 2.5 Flash76.4#467
7Gemini 3.7 Flash70.2#38

No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.

Interactive version: theaggregate.ai/benchmark?slug=verification-transport-planrewardbench-prb-v-llm-delegation-restoration-recovery · How It Works · Data refreshed daily, snapshot 2026-11-17.