Qwen3.5-2B

Alibaba2025-09-01
Overall Rank
#16
of 22 models
Overall Score
63.2
avg across benchmarks
Best Task
Key Information Extraction
78.5
Weakest Task
Text Extraction
37.9
Benchmark Performance
OlmOCR Benchv1.0
8/22
| Overall | Math | Table | Present | Absent | Order |
|---|---|---|---|---|---|
| 73.7 | 81.2 | 80.7 | 64.5 | 58.7 | 67.7 |
OmniDocBenchv1.5
18/22
| Overall | Text Edit↓ | CDM↑ | TEDS↑ | TEDS-S↑ | Read Order↓ |
|---|---|---|---|---|---|
| 48.7 | 0.621 | 62.9 | 45.3 | 48.2 | 0.401 |
IDP Core Benchv1.0
15/22
| Overall | KIE | OCR | Table | VQA |
|---|---|---|---|---|
| 67.1 | 78.5 | 56.2 | 72.4 | 59.8 |
Capability Profile
Strength Analysis
Auto-generated from benchmark scores
Strengths
- Key Information Extraction78.5
- Table Understanding66.1
Weaknesses
- Text Extraction37.9
- Visual QA59.8