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
84.75
GPT-5.6
B · Vision LLM
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 | GPT-5.6 | |
|---|---|---|
| Type | Vision LLM | Vision LLM |
| Maker | OpenAI | |
| Rank | 16 of 18 | 6 of 18 |
| F1 | 62.8 | 84.7 |
| Precision | 69.1 | 85.9 |
| Recall | 66.2 | 87.2 |
| Cost per 1,000 pages | $0.56 | $74 |
| Median time per resume | 29 s | 38 s |
| Resumes won | 24 | 273 |
Where each one wins
Where the two differ most.
| Kind of resume | # of resumes | Gemma 4 26B | GPT-5.6 | Gap |
|---|---|---|---|---|
| Scanned & rotated (source) | 19 | 40.7 | 71.6 | −30.8 |
| Rotated & warped pages (distractors) | 21 | 43.4 | 73.2 | −29.8 |
| Typos kept on purpose (distractors) | 37 | 51.3 | 80.6 | −29.3 |
| Hidden text & prompt injections (distractors) | 14 | 48.6 | 77.5 | −29.0 |
| Wording traps (distractors) | 16 | 60.6 | 89.1 | −28.5 |
| Distractor layouts (source) | 18 | 62.4 | 90.1 | −27.7 |
| Long bullet lists (extraction challenges) | 101 | 57.4 | 84.9 | −27.5 |
| Content as images (distractors) | 8 | 65.6 | 93.0 | −27.4 |
| Accented text (extraction challenges) | 13 | 59.3 | 86.5 | −27.3 |
| Many entries (extraction challenges) | 83 | 58.0 | 84.5 | −26.4 |
| Overlapping jobs (extraction challenges) | 69 | 57.5 | 83.2 | −25.7 |
| Promotions (extraction challenges) | 88 | 59.5 | 85.0 | −25.5 |
| Section | Resumes | Gemma 4 26B | GPT-5.6 | Gap |
|---|---|---|---|---|
| Contact details | 300 | 84.6 | 94.1 | −9.5 |
| Experience | 300 | 56.0 | 88.4 | −32.5 |
| Education | 297 | 68.0 | 91.9 | −23.9 |
| Summary | 290 | 35.5 | 83.4 | −47.9 |
| Certifications | 244 | 63.0 | 77.9 | −14.9 |
| Awards | 110 | 55.7 | 81.2 | −25.4 |
| Skills | 290 | 81.9 | 93.2 | −11.3 |
| Languages | 171 | 80.9 | 85.9 | −4.9 |
| Custom sections | 223 | 32.3 | 51.2 | −19.0 |
| Interests | 67 | 71.8 | 91.4 | −19.5 |
| Volunteering | 78 | 59.4 | 85.7 | −26.3 |
| Profiles | 37 | 57.2 | 69.4 | −12.2 |
| Projects | 83 | 55.0 | 88.2 | −33.2 |
| Publications | 53 | 54.5 | 77.8 | −23.3 |
Published resumes they disagree on most
Open one to see both outputs field by field.
| Resume | Gemma 4 26B | GPT-5.6 | Gap |
|---|---|---|---|
| Creative Director: Accented text in a styled CV | 39.3 | 90.8 | −51.5 |
| Supervisory Human Resources Specialist, GS-0201-15: Resume-builder template | 29.4 | 78.6 | −49.2 |
| Deputy Head of Mission: Rotated scan with hidden identity text | 34.1 | 76.2 | −42.1 |
| Head of Developer Experience (Internal title: Jedi Master of Analytics) : Unusual job titles | 53.7 | 93.6 | −39.9 |
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
- Gemma 4 26B vs LlamaParse Extract Agentic Plus
- Gemma 4 26B vs LlamaParse Extract Agentic
- Gemma 4 26B vs LlamaParse Extract Cost-Effective
- Gemma 4 26B vs Reducto Deep Extract
- Gemma 4 26B vs Reducto Extract
- Gemma 4 26B vs Extend Max Context
- Gemma 4 26B vs Extend Extract
- Gemma 4 26B vs GPT-5.5
- Gemma 4 26B vs GPT-5.4 Nano
- Gemma 4 26B vs Claude Opus 4.8
- Gemma 4 26B vs Gemini 3.5 Flash
- Gemma 4 26B vs Qwen 3.8 Flash
- Gemma 4 26B vs Qwen 3.8 27B
- Gemma 4 26B vs Qwen 3.5 35B
- Gemma 4 26B vs Qwen 3.5 9B
- Gemma 4 26B vs GLM 5.3 Flash
GPT-5.6 against every other model
- GPT-5.6 vs LlamaParse Extract Agentic Plus
- GPT-5.6 vs LlamaParse Extract Agentic
- GPT-5.6 vs LlamaParse Extract Cost-Effective
- GPT-5.6 vs Reducto Deep Extract
- GPT-5.6 vs Reducto Extract
- GPT-5.6 vs Extend Max Context
- GPT-5.6 vs Extend Extract
- GPT-5.6 vs GPT-5.5
- GPT-5.6 vs GPT-5.4 Nano
- GPT-5.6 vs Claude Opus 4.8
- GPT-5.6 vs Gemini 3.5 Flash
- GPT-5.6 vs Qwen 3.8 Flash
- GPT-5.6 vs Qwen 3.8 27B
- GPT-5.6 vs Qwen 3.5 35B
- GPT-5.6 vs Qwen 3.5 9B
- GPT-5.6 vs GLM 5.3 Flash
Popular comparisons
- LlamaParse Extract Agentic Plus vs Reducto Deep Extract
- Extend Max Context vs Reducto Deep Extract
- Extend Max Context vs LlamaParse Extract Agentic Plus
- Claude Opus 4.8 vs GPT-5.6
- Gemini 3.5 Flash vs GPT-5.6
- Claude Opus 4.8 vs Gemini 3.5 Flash
- LlamaParse Extract Agentic vs LlamaParse Extract Agentic Plus
- Reducto Deep Extract vs Reducto Extract
- Gemma 4 26B vs Qwen 3.8 27B
- Document APIs vs vision LLMs