GPT-5.5 vs Qwen 3.8 27B: which is better at resume extraction?
Qwen 3.8 27B leads GPT-5.5 by 0.3 points (83.8 vs 83.5 F1). Cost per 1,000 pages: $87 vs $9.59. Median time per resume: 48 s vs 36 s.
GPT-5.5 wins on 118 of 300 resumes, Qwen 3.8 27B on 139, 43 tied.
83.50
GPT-5.5
A · Vision LLM
83.77
Qwen 3.8 27B
B · Vision LLM
Qwen 3.8 27B is ahead by 0.27 points on these 300 resumes.
- Many entries 83A 83.8 · B 81.6+2.2
- Long bullet lists 101A 84.3 · B 80.8+3.5
- Split across pages 85A 85.0 · B 81.8+3.2
- Big skills list 60A 85.5 · B 82.9+2.6
- Custom sections 105A 80.8 · B 81.0−0.2
- Combined headings 37A 83.4 · B 83.7−0.3
- Promotions 88A 84.0 · B 83.2+0.7
- Partial dates 138A 84.0 · B 83.6+0.5
- Overlapping jobs 69A 82.1 · B 82.4−0.3
- Accented text 13A 83.9 · B 84.9−1.1
Side by side
| GPT-5.5 | Qwen 3.8 27B | |
|---|---|---|
| Type | Vision LLM | Vision LLM |
| Maker | OpenAI | Alibaba |
| Rank | 11 of 18 | 9 of 18 |
| F1 | 83.5 | 83.8 |
| Precision | 84.9 | 85.4 |
| Recall | 86.0 | 85.5 |
| Cost per 1,000 pages | $87 | $9.59 |
| Median time per resume | 48 s | 36 s |
| Resumes won | 118 | 139 |
Where each one wins
Where the two differ most.
| Kind of resume | # of resumes | GPT-5.5 | Qwen 3.8 27B | Gap |
|---|---|---|---|---|
| Scanned & rotated (source) | 19 | 60.7 | 78.4 | −17.7 |
| Rotated & warped pages (distractors) | 21 | 63.4 | 78.7 | −15.4 |
| Image only (reading the page) | 28 | 57.0 | 70.5 | −13.5 |
| Hidden text & prompt injections (distractors) | 14 | 69.1 | 82.4 | −13.3 |
| Typos kept on purpose (distractors) | 37 | 73.3 | 82.9 | −9.6 |
| Long (length) | 62 | 77.3 | 73.4 | +3.9 |
| References (noise) | 55 | 88.5 | 84.9 | +3.5 |
| Long bullet lists (extraction challenges) | 101 | 84.3 | 80.8 | +3.5 |
| Split across pages (extraction challenges) | 85 | 85.0 | 81.8 | +3.2 |
| Content as images (distractors) | 8 | 92.6 | 89.5 | +3.1 |
| Three-column (layout) | 20 | 82.6 | 85.5 | −2.9 |
| Resume-builder templates (source) | 10 | 83.9 | 86.7 | −2.8 |
| Section | Resumes | GPT-5.5 | Qwen 3.8 27B | Gap |
|---|---|---|---|---|
| Contact details | 300 | 93.5 | 93.2 | +0.3 |
| Experience | 300 | 87.4 | 86.2 | +1.2 |
| Education | 297 | 88.7 | 87.2 | +1.6 |
| Summary | 291 | 81.8 | 81.7 | +0.1 |
| Certifications | 254 | 77.3 | 82.7 | −5.4 |
| Awards | 98 | 75.3 | 78.4 | −3.2 |
| Skills | 287 | 91.5 | 92.7 | −1.2 |
| Languages | 168 | 85.7 | 85.4 | +0.3 |
| Custom sections | 210 | 49.7 | 51.2 | −1.5 |
| Interests | 64 | 92.4 | 94.6 | −2.2 |
| Volunteering | 77 | 85.0 | 84.7 | +0.3 |
| Profiles | 36 | 68.5 | 72.5 | −4.0 |
| Projects | 79 | 86.4 | 85.5 | +0.9 |
| Publications | 51 | 82.5 | 77.2 | +5.3 |
Published resumes they disagree on most
Open one to see both outputs field by field.
| Resume | GPT-5.5 | Qwen 3.8 27B | Gap |
|---|---|---|---|
| Head of Developer Experience (Internal title: Jedi Master of Analytics) : Unusual job titles | 80.2 | 96.3 | −16.2 |
| Lead Respiratory Therapist (Adult ICU): Long resume, 34 entries | 88.7 | 74.2 | +14.6 |
| Senior Controls Engineer: Five custom sections | 87.9 | 73.5 | +14.5 |
| Creative Director: Accented text in a styled CV | 90.2 | 78.3 | +11.9 |
Which to pick
Qwen 3.8 27B is ahead on accuracy, cost and speed.
GPT-5.5 does better on long, references and long bullet lists. Qwen 3.8 27B does better on scanned & rotated, rotated & warped pages and image only.
GPT-5.5 against every other model
- GPT-5.5 vs LlamaParse Extract Agentic Plus
- GPT-5.5 vs LlamaParse Extract Agentic
- GPT-5.5 vs LlamaParse Extract Cost-Effective
- GPT-5.5 vs Reducto Deep Extract
- GPT-5.5 vs Reducto Extract
- GPT-5.5 vs Extend Max Context
- GPT-5.5 vs Extend Extract
- GPT-5.5 vs GPT-5.6
- GPT-5.5 vs GPT-5.4 Nano
- GPT-5.5 vs Claude Opus 4.8
- GPT-5.5 vs Gemini 3.5 Flash
- GPT-5.5 vs Gemma 4 26B
- GPT-5.5 vs Qwen 3.8 Flash
- GPT-5.5 vs Qwen 3.5 35B
- GPT-5.5 vs Qwen 3.5 9B
- GPT-5.5 vs GLM 5.3 Flash
Qwen 3.8 27B against every other model
- Qwen 3.8 27B vs LlamaParse Extract Agentic Plus
- Qwen 3.8 27B vs LlamaParse Extract Agentic
- Qwen 3.8 27B vs LlamaParse Extract Cost-Effective
- Qwen 3.8 27B vs Reducto Deep Extract
- Qwen 3.8 27B vs Reducto Extract
- Qwen 3.8 27B vs Extend Max Context
- Qwen 3.8 27B vs Extend Extract
- Qwen 3.8 27B vs GPT-5.6
- Qwen 3.8 27B vs GPT-5.4 Nano
- Qwen 3.8 27B vs Claude Opus 4.8
- Qwen 3.8 27B vs Gemini 3.5 Flash
- Qwen 3.8 27B vs Gemma 4 26B
- Qwen 3.8 27B vs Qwen 3.8 Flash
- Qwen 3.8 27B vs Qwen 3.5 35B
- Qwen 3.8 27B vs Qwen 3.5 9B
- Qwen 3.8 27B 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