Skip to content
vs

GPT-5.4 Nano vs Qwen 3.8 27B: which is better at resume extraction?

Qwen 3.8 27B leads GPT-5.4 Nano by 21.9 points (83.8 vs 61.9 F1). Cost per 1,000 pages: $2.76 vs $9.59. Median time per resume: 20 s vs 36 s.

GPT-5.4 Nano wins on 25 of 300 resumes, Qwen 3.8 27B on 272, 3 tied.

61.86

GPT-5.4 Nano

A · Vision LLM

−21.91A − B, points of F1

83.77

Qwen 3.8 27B

B · Vision LLM

A better on 25level on 3B better on 272

Qwen 3.8 27B is ahead by 21.91 points on these 300 resumes.

  • Many entries 83A 59.7 · B 81.6−22.0
  • Long bullet lists 101A 61.0 · B 80.8−19.8
  • Split across pages 85A 61.7 · B 81.8−20.1
  • Big skills list 60A 64.9 · B 82.9−18.0
  • Custom sections 105A 59.5 · B 81.0−21.5
  • Combined headings 37A 64.6 · B 83.7−19.1
  • Promotions 88A 58.7 · B 83.2−24.5
  • Partial dates 138A 62.6 · B 83.6−21.0
  • Overlapping jobs 69A 58.6 · B 82.4−23.7
  • Accented text 13A 55.9 · B 84.9−29.1

Side by side

GPT-5.4 Nano vs Qwen 3.8 27B
GPT-5.4 NanoQwen 3.8 27B
TypeVision LLMVision LLM
MakerOpenAIAlibaba
Rank17 of 189 of 18
F161.983.8
Precision70.185.4
Recall66.385.5
Cost per 1,000 pages$2.76$9.59
Median time per resume20 s36 s
Resumes won25272

GPT-5.4 Nano in detailQwen 3.8 27B in detail

Where each one wins

Where the two differ most.

GPT-5.4 Nano vs Qwen 3.8 27B by kind of resume
Kind of resume# of resumesGPT-5.4 NanoQwen 3.8 27BGap
Scanned & rotated (source)196.378.4−72.1
Rotated & warped pages (distractors)2111.778.7−67.1
Hidden text & prompt injections (distractors)1422.082.4−60.4
Image only (reading the page)2821.070.5−49.5
Typos kept on purpose (distractors)3733.582.9−49.4
Two-column sidebar (layout)5447.478.4−31.0
Accented text (extraction challenges)1355.984.9−29.1
Content as images (distractors)861.389.5−28.2
Sidebar or grid (reading the page)8254.981.8−26.9
Distractor layouts (source)1862.287.6−25.4
Medium (length)9060.285.5−25.3
Wording traps (distractors)1662.887.6−24.8
GPT-5.4 Nano vs Qwen 3.8 27B by section
SectionResumesGPT-5.4 NanoQwen 3.8 27BGap
Contact details30081.293.2−12.0
Experience30063.786.2−22.5
Education29762.287.2−25.0
Summary29071.781.7−10.0
Certifications24453.682.7−29.2
Awards10244.978.4−33.5
Skills29174.892.7−17.9
Languages17069.085.4−16.4
Custom sections23126.351.2−24.9
Interests6670.094.6−24.6
Volunteering8059.984.7−24.8
Profiles6623.772.5−48.8
Projects8148.885.5−36.7
Publications5130.777.2−46.5

Published resumes they disagree on most

Open one to see both outputs field by field.

Which to pick

Qwen 3.8 27B is more accurate. GPT-5.4 Nano costs 71% less for 21.9 fewer points.

Qwen 3.8 27B does better on scanned & rotated, rotated & warped pages and hidden text & prompt injections.

GPT-5.4 Nano against every other model

Qwen 3.8 27B against every other model

Popular comparisons

Frequently asked questions (FAQs)

Which is better for resume extraction, GPT-5.4 Nano or Qwen 3.8 27B?

Qwen 3.8 27B: 83.8 F1 against 61.9.

Which is cheaper, GPT-5.4 Nano or Qwen 3.8 27B?

GPT-5.4 Nano: $2.76 per 1,000 pages against $9.59.

Where does each one do better?

Qwen 3.8 27B: scanned & rotated, rotated & warped pages and hidden text & prompt injections.