Tihu benchmark sample 1 at medium effort, dated by the source data day. The upstream AI reference score is 8.51 ± 0.12.
Date evidence, prompt & provenance
Date precision / evidence: day / source-benchmark-data-date. Recorded date: 2026-10-13; source publication / generation date not separately verified. Generation date is not independently authenticated.
Data registrada
2026-10-13
Date precision
Day
Date basis
source-benchmark-data-date
Modelo informado pela fonte
GPT-6.1 Sol
Model attribution
Source-reported, not independently authenticated
Autor / publicador
Tihu · 鹈鹕测试榜
Formato
SVG estático
Categoria do prompt
Source-documented test
Prompt original
Generate an SVG of a pelican riding a bicycle
Last updated
2026-11-10
Original catalog label: medium. A historical organizing label, not a score from a controlled benchmark.
Independent sample 01 for this model in Sergio Paniego’s public OpenEnv pelican SVG dataset. The model label, provider and sample number follow its per-row metadata. This is the PNG published by the dataset, not an image regenerated here. The date is the dataset publication date, not the exact generation time. Model provenance and environment scores are not independently authenticated here, and scores are not presented as capability rankings.
Independent sample 02 for this model in Sergio Paniego’s public OpenEnv pelican SVG dataset. The model label, provider and sample number follow its per-row metadata. This is the PNG published by the dataset, not an image regenerated here. The date is the dataset publication date, not the exact generation time. Model provenance and environment scores are not independently authenticated here, and scores are not presented as capability rankings.
Independent sample 03 for this model in Sergio Paniego’s public OpenEnv pelican SVG dataset. The model label, provider and sample number follow its per-row metadata. This is the PNG published by the dataset, not an image regenerated here. The date is the dataset publication date, not the exact generation time. Model provenance and environment scores are not independently authenticated here, and scores are not presented as capability rankings.