Skip to content
vs

Gemma 4 26B vs GPT-5.6: which is better at resume extraction?

GPT-5.6 leads Gemma 4 26B by 21.9 points (84.7 vs 62.8 F1). Cost per 1,000 pages: $0.56 vs $74. Median time per resume: 29 s vs 38 s.

Gemma 4 26B wins on 24 of 300 resumes, GPT-5.6 on 273, 3 tied.

62.82

Gemma 4 26B

A · Vision LLM

−21.93A − B, points of F1

84.75

GPT-5.6

B · Vision LLM

A better on 24level on 3B better on 273

GPT-5.6 is ahead by 21.93 points on these 300 resumes.

  • Many entries 83A 58.0 · B 84.5−26.4
  • Long bullet lists 101A 57.4 · B 84.9−27.5
  • Split across pages 85A 60.0 · B 84.9−24.9
  • Big skills list 60A 61.9 · B 85.2−23.2
  • Custom sections 105A 58.9 · B 82.2−23.2
  • Combined headings 37A 60.9 · B 84.1−23.2
  • Promotions 88A 59.5 · B 85.0−25.5
  • Partial dates 138A 62.9 · B 85.7−22.7
  • Overlapping jobs 69A 57.5 · B 83.2−25.7
  • Accented text 13A 59.3 · B 86.5−27.3

Side by side

Gemma 4 26B vs GPT-5.6
Gemma 4 26BGPT-5.6
TypeVision LLMVision LLM
MakerGoogleOpenAI
Rank16 of 186 of 18
F162.884.7
Precision69.185.9
Recall66.287.2
Cost per 1,000 pages$0.56$74
Median time per resume29 s38 s
Resumes won24273

Gemma 4 26B in detailGPT-5.6 in detail

Where each one wins

Where the two differ most.

Gemma 4 26B vs GPT-5.6 by kind of resume
Kind of resume# of resumesGemma 4 26BGPT-5.6Gap
Scanned & rotated (source)1940.771.6−30.8
Rotated & warped pages (distractors)2143.473.2−29.8
Typos kept on purpose (distractors)3751.380.6−29.3
Hidden text & prompt injections (distractors)1448.677.5−29.0
Wording traps (distractors)1660.689.1−28.5
Distractor layouts (source)1862.490.1−27.7
Long bullet lists (extraction challenges)10157.484.9−27.5
Content as images (distractors)865.693.0−27.4
Accented text (extraction challenges)1359.386.5−27.3
Many entries (extraction challenges)8358.084.5−26.4
Overlapping jobs (extraction challenges)6957.583.2−25.7
Promotions (extraction challenges)8859.585.0−25.5
Gemma 4 26B vs GPT-5.6 by section
SectionResumesGemma 4 26BGPT-5.6Gap
Contact details30084.694.1−9.5
Experience30056.088.4−32.5
Education29768.091.9−23.9
Summary29035.583.4−47.9
Certifications24463.077.9−14.9
Awards11055.781.2−25.4
Skills29081.993.2−11.3
Languages17180.985.9−4.9
Custom sections22332.351.2−19.0
Interests6771.891.4−19.5
Volunteering7859.485.7−26.3
Profiles3757.269.4−12.2
Projects8355.088.2−33.2
Publications5354.577.8−23.3

Published resumes they disagree on most

Open one to see both outputs field by field.

Which to pick

GPT-5.6 is more accurate. Gemma 4 26B costs 99% less for 21.9 fewer points.

GPT-5.6 does better on scanned & rotated, rotated & warped pages and typos kept on purpose.

Gemma 4 26B against every other model

GPT-5.6 against every other model

Popular comparisons

Frequently asked questions (FAQs)

Which is better for resume extraction, Gemma 4 26B or GPT-5.6?

GPT-5.6: 84.7 F1 against 62.8.

Which is cheaper, Gemma 4 26B or GPT-5.6?

Gemma 4 26B: $0.56 per 1,000 pages against $74.

Where does each one do better?

GPT-5.6: scanned & rotated, rotated & warped pages and typos kept on purpose.