Gemini 3.5 Flash vs GPT-5.6: which is better at resume extraction?
Gemini 3.5 Flash leads GPT-5.6 by 0.7 points (85.5 vs 84.7 F1). Cost per 1,000 pages: $35 vs $74. Median time per resume: 43 s vs 38 s.
Gemini 3.5 Flash wins on 133 of 300 resumes, GPT-5.6 on 121, 46 tied.
85.45
Gemini 3.5 Flash
A · Vision LLM
84.75
GPT-5.6
B · Vision LLM
Gemini 3.5 Flash is ahead by 0.70 points on these 300 resumes.
- Many entries 83A 84.3 · B 84.5−0.2
- Long bullet lists 101A 83.9 · B 84.9−1.1
- Split across pages 85A 84.6 · B 84.9−0.3
- Big skills list 60A 86.0 · B 85.2+0.9
- Custom sections 105A 83.6 · B 82.2+1.4
- Combined headings 37A 84.2 · B 84.1+0.2
- Promotions 88A 87.3 · B 85.0+2.4
- Partial dates 138A 84.9 · B 85.7−0.7
- Overlapping jobs 69A 86.4 · B 83.2+3.3
- Accented text 13A 86.4 · B 86.5−0.2
Side by side
| Gemini 3.5 Flash | GPT-5.6 | |
|---|---|---|
| Type | Vision LLM | Vision LLM |
| Maker | OpenAI | |
| Rank | 5 of 18 | 6 of 18 |
| F1 | 85.5 | 84.7 |
| Precision | 86.9 | 85.9 |
| Recall | 87.9 | 87.2 |
| Cost per 1,000 pages | $35 | $74 |
| Median time per resume | 43 s | 38 s |
| Resumes won | 133 | 121 |
Where each one wins
Where the two differ most.
| Kind of resume | # of resumes | Gemini 3.5 Flash | GPT-5.6 | Gap |
|---|---|---|---|---|
| Content as images (distractors) | 8 | 76.7 | 93.0 | −16.3 |
| Scanned & rotated (source) | 19 | 87.2 | 71.6 | +15.7 |
| Rotated & warped pages (distractors) | 21 | 87.4 | 73.2 | +14.2 |
| Image only (reading the page) | 28 | 75.1 | 65.4 | +9.7 |
| Hidden text & prompt injections (distractors) | 14 | 86.7 | 77.5 | +9.1 |
| Wording traps (distractors) | 16 | 80.4 | 89.1 | −8.7 |
| Distractor layouts (source) | 18 | 81.4 | 90.1 | −8.7 |
| Two-column sidebar (layout) | 54 | 83.4 | 78.3 | +5.1 |
| Typos kept on purpose (distractors) | 37 | 84.4 | 80.6 | +3.8 |
| Sidebar or grid (reading the page) | 82 | 85.1 | 81.7 | +3.4 |
| Overlapping jobs (extraction challenges) | 69 | 86.4 | 83.2 | +3.3 |
| Busy page (noise) | 18 | 80.8 | 84.0 | −3.2 |
| Section | Resumes | Gemini 3.5 Flash | GPT-5.6 | Gap |
|---|---|---|---|---|
| Contact details | 300 | 95.3 | 94.1 | +1.2 |
| Experience | 300 | 86.5 | 88.4 | −1.9 |
| Education | 297 | 91.5 | 91.9 | −0.4 |
| Summary | 290 | 89.3 | 83.4 | +5.9 |
| Certifications | 245 | 84.0 | 77.9 | +6.1 |
| Awards | 92 | 80.3 | 81.2 | −0.9 |
| Skills | 289 | 93.5 | 93.2 | +0.2 |
| Languages | 170 | 84.3 | 85.9 | −1.5 |
| Custom sections | 216 | 50.1 | 51.2 | −1.1 |
| Interests | 63 | 94.2 | 91.4 | +2.8 |
| Volunteering | 79 | 83.7 | 85.7 | −1.9 |
| Profiles | 35 | 64.8 | 69.4 | −4.6 |
| Projects | 80 | 80.4 | 88.2 | −7.8 |
| Publications | 52 | 83.4 | 77.8 | +5.6 |
Published resumes they disagree on most
Open one to see both outputs field by field.
| Resume | Gemini 3.5 Flash | GPT-5.6 | Gap |
|---|---|---|---|
| Senior Data Scientist (Lead): Fifty skills, seven pages | 95.8 | 72.8 | +23.0 |
| Senior Game Designer: Game designer's Canva resume | 54.8 | 75.9 | −21.1 |
| Creative Director: Accented text in a styled CV | 74.7 | 90.8 | −16.0 |
| Head of Events & Public Space Operations: Table grid with emoji bullets | 76.7 | 90.0 | −13.3 |
Which to pick
Gemini 3.5 Flash is more accurate. GPT-5.6 is the faster of the two.
Gemini 3.5 Flash does better on scanned & rotated, rotated & warped pages and image only. GPT-5.6 does better on content as images, wording traps and distractor layouts.
Gemini 3.5 Flash against every other model
- Gemini 3.5 Flash vs LlamaParse Extract Agentic Plus
- Gemini 3.5 Flash vs LlamaParse Extract Agentic
- Gemini 3.5 Flash vs LlamaParse Extract Cost-Effective
- Gemini 3.5 Flash vs Reducto Deep Extract
- Gemini 3.5 Flash vs Reducto Extract
- Gemini 3.5 Flash vs Extend Max Context
- Gemini 3.5 Flash vs Extend Extract
- Gemini 3.5 Flash vs GPT-5.5
- Gemini 3.5 Flash vs GPT-5.4 Nano
- Gemini 3.5 Flash vs Claude Opus 4.8
- Gemini 3.5 Flash vs Gemma 4 26B
- Gemini 3.5 Flash vs Qwen 3.8 Flash
- Gemini 3.5 Flash vs Qwen 3.8 27B
- Gemini 3.5 Flash vs Qwen 3.5 35B
- Gemini 3.5 Flash vs Qwen 3.5 9B
- Gemini 3.5 Flash 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 Gemma 4 26B
- 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
- 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