PCFBench - Material Rate Extraction: leaderboard
Metric: Claim F1 (%; extracting material-emission-rate claims with evidence from source PDFs, 22 questions and 55 claims; from PCFBench, a product-carbon-footprint benchmark of 614 expert-labelled items; scored deterministically or by a temperature-0 judge; higher is better). Source: arxiv.org. Saturation forecast: Around 2033. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V3.2 | 51 |
| 2 | Gemini 3 Flash | 47 |
| 3 | Gemini 3.1 Pro (Preview) | 44 |
| 4 | Claude Sonnet 4.6 | 39 |
| 5 | Claude Haiku 4.5 | 36 |
| 6 | GPT-5.5 | 31 |
| 7 | GPT-5.4 Mini | 29 |
| 8 | Claude Opus 4.6 | 27 |
Interactive version: theaggregate.ai/benchmark?slug=pcfbench-material-rate-extraction · How It Works · Data refreshed daily, snapshot 2026-09-26.