Skip to content
vs

Claude Opus 4.8 vs GPT-5.5: which is better at resume extraction?

Claude Opus 4.8 leads GPT-5.5 by 2.0 points (85.5 vs 83.5 F1). Cost per 1,000 pages: $90 vs $87. Median time per resume: 31 s vs 48 s.

Claude Opus 4.8 wins on 139 of 300 resumes, GPT-5.5 on 113, 48 tied.

85.51

Claude Opus 4.8

A · Vision LLM

+2.00A − B, points of F1

83.50

GPT-5.5

B · Vision LLM

A better on 139level on 48B better on 113

Claude Opus 4.8 is ahead by 2.00 points on these 300 resumes.

  • Many entries 83A 83.7 · B 83.8−0.1
  • Long bullet lists 101A 83.1 · B 84.3−1.2
  • Split across pages 85A 84.6 · B 85.0−0.5
  • Big skills list 60A 84.8 · B 85.5−0.7
  • Custom sections 105A 83.4 · B 80.8+2.6
  • Combined headings 37A 85.1 · B 83.4+1.7
  • Promotions 88A 85.5 · B 84.0+1.5
  • Partial dates 138A 84.5 · B 84.0+0.5
  • Overlapping jobs 69A 83.5 · B 82.1+1.4
  • Accented text 13A 85.6 · B 83.9+1.7

Side by side

Claude Opus 4.8 vs GPT-5.5
Claude Opus 4.8GPT-5.5
TypeVision LLMVision LLM
MakerAnthropicOpenAI
Rank4 of 1811 of 18
F185.583.5
Precision86.884.9
Recall87.586.0
Cost per 1,000 pages$90$87
Median time per resume31 s48 s
Resumes won139113

Claude Opus 4.8 in detailGPT-5.5 in detail

Where each one wins

Where the two differ most.

Claude Opus 4.8 vs GPT-5.5 by kind of resume
Kind of resume# of resumesClaude Opus 4.8GPT-5.5Gap
Scanned & rotated (source)1987.560.7+26.8
Rotated & warped pages (distractors)2187.463.4+24.0
Hidden text & prompt injections (distractors)1490.869.1+21.7
Image only (reading the page)2877.057.0+20.0
Typos kept on purpose (distractors)3785.273.3+11.9
Content as images (distractors)883.692.6−8.9
Two-column sidebar (layout)5481.977.2+4.6
Sidebar or grid (reading the page)8284.280.4+3.8
Distractor layouts (source)1882.886.6−3.8
Handwritten (reading the page)954.751.0+3.7
Three-column (layout)2086.282.6+3.6
Wording traps (distractors)1681.985.3−3.3
Claude Opus 4.8 vs GPT-5.5 by section
SectionResumesClaude Opus 4.8GPT-5.5Gap
Contact details30093.793.5+0.3
Experience30088.287.4+0.8
Education29790.288.7+1.5
Summary29089.781.8+7.9
Certifications23985.477.3+8.0
Awards10672.975.3−2.3
Skills28395.291.5+3.7
Languages16785.885.7+0.1
Custom sections21851.649.7+1.9
Interests6393.792.4+1.3
Volunteering7884.685.0−0.4
Profiles3561.468.5−7.1
Projects7882.186.4−4.3
Publications5281.882.5−0.7

Published resumes they disagree on most

Open one to see both outputs field by field.

Which to pick

Claude Opus 4.8 is more accurate. GPT-5.5 costs 3% less for 2.0 fewer points.

Claude Opus 4.8 does better on scanned & rotated, rotated & warped pages and hidden text & prompt injections. GPT-5.5 does better on content as images, distractor layouts and wording traps.

Claude Opus 4.8 against every other model

GPT-5.5 against every other model

Popular comparisons

Frequently asked questions (FAQs)

Which is better for resume extraction, Claude Opus 4.8 or GPT-5.5?

Claude Opus 4.8: 85.5 F1 against 83.5.

Which is cheaper, Claude Opus 4.8 or GPT-5.5?

GPT-5.5: $87 per 1,000 pages against $90.

Where does each one do better?

Claude Opus 4.8: scanned & rotated, rotated & warped pages and hidden text & prompt injections. GPT-5.5: content as images, distractor layouts and wording traps.