IndustryBench - Process Principles: leaderboard

IndustryBench (2026) restricted to its 528 Process Principles questions (process flow and parameter-outcome relationships), Chinese industrial procurement QA answered closed-book in default non-reasoning mode and scored 0 to 3 by a Qwen3-Max judge, with answers that violate a safety clause set to 0. Mean score out of 3; higher is better.

Metric: Final (SV) score (out of 3): mean 0-3 rubric score from a Qwen3-Max judge against the reference answer, set to 0 for any answer with a safety violation against the source standard; closed-book, zero-shot, provider-default decoding, over the 528 items of the 2,049 Chinese industrial procurement QA items of IndustryBench grounded in Chinese national standards (GB/T) in the Process Principles capability dimension (process flow and parameter-outcome relationships; default non-reasoning runs); higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 17 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Qwen 3.6 PlusNon-reasoning2.46#238
2GPT-5.4Non-reasoning2.4#135
3Qwen 3 MaxNon-reasoning2.38#415
4Gemini 3.1 Pro (Preview)2.36#99
5GPT-5.2Non-reasoning2.35#185
6Qwen 3.5 122B A10BNon-reasoning2.31#346
7Qwen 3.5 PlusNon-reasoning2.31#242
8Claude Opus 4.62.28#113
9Qwen 3.5 397B A17BNon-reasoning2.28#264
10Qwen 3.5 27BNon-reasoning2.27#375
11Qwen 3.5 35B A3BNon-reasoning2.2#492
12Kimi K2.5Non-reasoning2.17#266
13GLM-5Non-reasoning2.11#269
14Claude Sonnet 4.62.08#173
15MiniMax-M2.52.07#600

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

Interactive version: theaggregate.ai/benchmark?slug=industrybench-process-principles · How It Works · Data refreshed daily, snapshot 2026-11-21.