Tihu · Pelican Bench

Claude Opus 5.5

Anthropic · Model release:

AI reference score 7.88 ± 0.51 · high (best) Last 4 weekly runs · 12 images
Tier comparison
TierSamplesAI reference scoreValid SVG rateMean timeSamples
high
Claude Opus 5.5 · high · 2026-10-25 · #3 · 8.39
 7.88 ± 0.51Last 4 weekly runs · 12 images 100%68 s3
medium default
Claude Opus 5.5 · medium · 2026-10-25 · #3 · 8.03
 7.86 ± 0.57Last 4 weekly runs · 12 images 100%49 s3
low
Claude Opus 5.5 · low · 2026-10-25 · #1 · 6.99
 6.38 ± 0.89Last 4 weekly runs · 12 images 100%29 s3
Claude Opus 5.5 · high · 2026-10-25 · #3 · 8.39
high 7.88 ± 0.51Last 4 weekly runs · 12 images
Valid SVG rate 100% · Mean time 68 s
Claude Opus 5.5 · medium · 2026-10-25 · #3 · 8.03
medium default7.86 ± 0.57Last 4 weekly runs · 12 images
Valid SVG rate 100% · Mean time 49 s
Claude Opus 5.5 · low · 2026-10-25 · #1 · 6.99
low 6.38 ± 0.89Last 4 weekly runs · 12 images
Valid SVG rate 100% · Mean time 29 s

Latest samples

high

AI reference score 7.88 ± 0.51 Last 4 weekly runs · 12 images Monitoring
Top tiers of one modelView tier page

SunLatest

Claude Opus 5.5 · high · 2026-10-25 · #3 · 8.39Claude Opus 5.5 · high · 2026-10-25 · #1 · 8.22Claude Opus 5.5 · high · 2026-10-25 · #2 · 7.21

medium

AI reference score 7.86 ± 0.57 Last 4 weekly runs · 12 images Monitoring
Top tiers of one modelView tier page

SunLatest

Claude Opus 5.5 · medium · 2026-10-25 · #3 · 8.03Claude Opus 5.5 · medium · 2026-10-25 · #2 · 7.69Claude Opus 5.5 · medium · 2026-10-25 · #1 · 7.63

low

AI reference score 6.38 ± 0.89 Last 4 weekly runs · 12 images Monitoring
Top tiers of one modelView tier page

SunLatest

Claude Opus 5.5 · low · 2026-10-25 · #1 · 6.99Claude Opus 5.5 · low · 2026-10-25 · #3 · 5.84Claude Opus 5.5 · low · 2026-10-25 · #2 · 5.46

Score trend

high Normal

2026-10-08 · 8.022026-10-09 · 8.022026-10-10 · 8.022026-10-11 · 8.092026-10-12 · 8.092026-10-13 · 8.092026-10-14 · 8.092026-10-15 · 8.092026-10-16 · 8.092026-10-17 · 8.092026-10-18 · 7.482026-10-19 · 7.482026-10-20 · 7.482026-10-21 · 7.482026-10-23 · 7.482026-10-24 · 7.482026-10-25 · 7.942026-10-26 · 7.94

18 records in the last 30 days

2026-10-26 · Normal2026-10-25 · Normal2026-10-24 · Normal2026-10-23 · Normal2026-10-21 · Normal2026-10-20 · Normal2026-10-19 · Normal2026-10-18 · Collecting earlier results2026-10-17 · Normal2026-10-16 · Normal2026-10-15 · Normal2026-10-14 · Normal2026-10-13 · Normal2026-10-12 · Normal2026-10-11 · Collecting earlier results2026-10-10 · Normal2026-10-09 · Normal2026-10-08 · Collecting earlier results

medium Normal

2026-10-08 · 7.672026-10-09 · 7.672026-10-10 · 7.672026-10-11 · 8.232026-10-12 · 8.232026-10-13 · 8.232026-10-14 · 8.232026-10-15 · 8.232026-10-16 · 8.232026-10-17 · 8.232026-10-18 · 7.772026-10-19 · 7.772026-10-20 · 7.772026-10-21 · 7.772026-10-23 · 7.772026-10-24 · 7.772026-10-25 · 7.782026-10-26 · 7.78

18 records in the last 30 days

2026-10-26 · Normal2026-10-25 · Normal2026-10-24 · Normal2026-10-23 · Normal2026-10-21 · Normal2026-10-20 · Normal2026-10-19 · Normal2026-10-18 · Collecting earlier results2026-10-17 · Normal2026-10-16 · Normal2026-10-15 · Normal2026-10-14 · Normal2026-10-13 · Normal2026-10-12 · Normal2026-10-11 · Collecting earlier results2026-10-10 · Normal2026-10-09 · Normal2026-10-08 · Collecting earlier results

low Normal

2026-10-08 · 5.942026-10-09 · 5.942026-10-10 · 5.942026-10-11 · 7.452026-10-12 · 7.452026-10-13 · 7.452026-10-14 · 7.452026-10-15 · 7.452026-10-16 · 7.452026-10-17 · 7.452026-10-18 · 6.062026-10-19 · 6.062026-10-20 · 6.062026-10-21 · 6.062026-10-23 · 6.062026-10-24 · 6.062026-10-25 · 6.102026-10-26 · 6.10

18 records in the last 30 days

2026-10-26 · Normal2026-10-25 · Normal2026-10-24 · Normal2026-10-23 · Normal2026-10-21 · Normal2026-10-20 · Normal2026-10-19 · Normal2026-10-18 · Collecting earlier results2026-10-17 · Normal2026-10-16 · Normal2026-10-15 · Normal2026-10-14 · Normal2026-10-13 · Normal2026-10-12 · Normal2026-10-11 · Collecting earlier results2026-10-10 · Normal2026-10-09 · Normal2026-10-08 · Collecting earlier results

Earlier samples

high

medium

low

Versions by score

  1. Claude Opus 5.5
    Claude Opus 5.5 · high
    high 7.88Last 4 weekly runs · 12 images

Scores by criterion

Scores by tier
Scoring criteria high medium low
Pelican likeness8.678.757.58
Bicycle correctness7.757.835.83
Riding pose7.006.674.33
Composition8.838.427.00
Obvious errors (higher is better)7.426.925.50

Release notes

Original (Chinese)官方文档标注 Released September 22, 2026；当天在 Claude 应用、Claude API（claude-opus-5-5）及 AWS/Google Cloud/Microsoft Foundry 可用。

· View release source ↗