Qwen3.5-4B

Alibaba2025-09-01
Overall Rank
#10
of 22 models
Overall Score
73.1
avg across benchmarks
Best Task
Key Information Extraction
86.0
Weakest Task
Text Extraction
70.8
Benchmark Performance
OlmOCR Benchv1.0
4/22
| Overall | Math | Table | Present | Absent | Order |
|---|---|---|---|---|---|
| 77.2 | 86.0 | 85.0 | 68.9 | 55.7 | 74.5 |
OmniDocBenchv1.5
16/22
| Overall | Text Edit↓ | CDM↑ | TEDS↑ | TEDS-S↑ | Read Order↓ |
|---|---|---|---|---|---|
| 67.6 | 0.292 | 71.5 | 60.4 | 64.6 | 0.106 |
IDP Core Benchv1.0
10/22
| Overall | KIE | OCR | Table | VQA |
|---|---|---|---|---|
| 74.5 | 86.0 | 64.7 | 76.7 | 72.4 |
Capability Profile
Strength Analysis
Auto-generated from benchmark scores
Strengths
- Key Information Extraction86.0
- Layout & Order82.0
Weaknesses
- Text Extraction70.8
- Formula71.5