Tihu · Pelican Bench

Hy3

Tencent · Model release:

AI reference score 6.33 ± 0.86 · default (best) Last 3 weekly runs · 9 images
Tier comparison
TierSamplesAI reference scoreValid SVG rateMean timeSamples
default · This model version has no selectable reasoning tiers
Hy3 · default · 2026-10-25 · #1 · 6.86
 6.33 ± 0.86Last 3 weekly runs · 9 images 100%110 s3
Hy3 · default · 2026-10-25 · #1 · 6.86
default · This model version has no selectable reasoning tiers 6.33 ± 0.86Last 3 weekly runs · 9 images
Valid SVG rate 100% · Mean time 110 s

Latest samples

default · This model version has no selectable reasoning tiers

AI reference score 6.33 ± 0.86 Last 3 weekly runs · 9 images 1 weekly test left
Top tiers of one modelView tier page

SunLatest

Hy3 · default · 2026-10-25 · #1 · 6.86Hy3 · default · 2026-10-25 · #3 · 5.78Hy3 · default · 2026-10-25 · #2 · 5.20

Score trend

default · This model version has no selectable reasoning tiers Normal

2026-10-11 · 6.092026-10-18 · 6.942026-10-19 · 6.942026-10-20 · 6.942026-10-21 · 6.942026-10-23 · 6.942026-10-24 · 6.942026-10-25 · 5.952026-10-26 · 5.95

9 records in the last 30 days

2026-10-26 · Normal2026-10-25 · Collecting earlier results2026-10-24 · Normal2026-10-23 · Normal2026-10-21 · Normal2026-10-20 · Normal2026-10-19 · Normal2026-10-18 · Collecting earlier results2026-10-11 · Collecting earlier results

Earlier samples

default · This model version has no selectable reasoning tiers

Versions by score

  1. Hy3
    Hy3 · default
    default · This model version has no selectable reasoning tiers 6.33Last 3 weekly runs · 9 images

Scores by criterion

Scores by tier
Scoring criteria default · This model version has no selectable reasoning tiers
Pelican likeness6.22
Bicycle correctness5.94
Riding pose5.78
Composition6.22
Obvious errors (higher is better)5.44

Release notes

Original (Chinese)当天 Hy3 可用。

· View release source ↗