LiveClin: A Live Clinical Benchmark for Vision-Language Models

LiveClin is a contamination-free, biannually updated clinical benchmark for evaluating large vision-language models. It was created by AQ-MedAI and last updated on February 27, 2026. The benchmark presents realistic, multi-stage clinical case reasoning with medical images and tables.

Use Cases

• Benchmarking model performance on clinical reasoning based on the described multi-stage diagnostic workflow.
• Evaluating multimodal model capabilities based on the described combination of medical images and tables.
• Assessing model robustness on contamination-free test sets based on the benchmark's stated design principle.

Strengths

• Designed to be contamination-free, a specific feature for reliable evaluation.
• Biannually updated, suggesting a commitment to maintaining relevance.
• Based on realistic clinical scenarios with a progressive diagnostic workflow.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

AQ-MedAI

Collection Method

Likely curated for benchmarking, as described.

Time Range

Biannually updated, but specific temporal coverage is unknown.

Freshness

Last updated 2026-02-27 13:02:01; freshness should be verified.

Geography

Region 'us' is indicated in platform tags, but spatial coverage is not explicitly defined.