Completed / CodeSOTA execution on White
Inspect the measurement.
Qwen/Qwen2-VL-2B-Instruct · OCRBench v1 / full serial reference · 78.5% accuracy
Full benchmark split · 1000 scored examples. All 1,000 OCRBench v1 questions. Upstream answer-containment scoring; response verbosity can game this metric. Separate visual-text QA protocol, not document parsing/layout or table reconstruction. Image/decoding budgets and model chat template are CodeSOTA-pinned, not a claim of vendor inference parity.
Run ID / completed
3f6b2514a95d7e1f9decf62a
18 Oct 2026, 18:33 UTC
18 Oct 2026, 18:33 UTC
Model revision
895c3a49bc3fa70a340399125c650a463535e71cComparison scope
dcccfe2d1c65816985fdcc2bb76bfad20fcc76719fc61e7a4fd8a82df9ec1d29Evaluator SHA-256
50e622b9a313b1ee83d79c785095f871b4e136fedc03ef0a73e48ac22c46cce5Observed resources
768s elapsed · 9.37 GiB peak torch allocation
Dollar cost: not measured
Dollar cost: not measured
Download evidence
Inspect individual examples
Read the recorded prediction, reference and outcome for each example.
- Scored examples
- 1,000
- Passed
- 785
- Failed
- 215
- Metric
- accuracy
Examples are downloaded only when requested. Search covers all loaded records; 25 examples are displayed per page. Download all examples · JSONL
Example evidence has not been loaded.