HPR-Bench - Tag Inference - Household Context: leaderboard

Metric: Avg@10 Tag Accuracy (%; mean of 10 samples). Source: arxiv.org. Saturation forecast: Around September 2027. 15 models tracked.

Top models

#ModelScore
1Qwen 3.6 Plus (Thinking)59.65
2Qwen 3.5 397B A17B (Thinking)54.66
3Qwen 3.5 Plus (Thinking)54
4Kimi K2 (Thinking)53.8
5Kimi K2.5 (Thinking)53.8
6GLM-5 (Thinking)53.45
7GLM-4.7 (Thinking)52
8DeepSeek V3.2 (Thinking)49.81
9DeepSeek R1 052847.52
10Qwen 3 235B A22B 2507 Instruct41.4
11Qwen 3 235B A22B 2507 (Thinking)41.35
12Qwen 3 8B (Thinking)38.31
13Qwen 3 8B (Non-reasoning)32.81

Interactive version: theaggregate.ai/benchmark?slug=hpr-bench-tag-inference-household-context · How It Works · Data refreshed daily, snapshot 2026-09-25.