Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Document-based Question Answering on Human Evaluation 40 document-query pairs
Loading...
4.4
Relevance
QASC
3.152
3.476
3.8
4.124
Apr 29, 2026
Relevance
Coherence
Completeness
Answer Quality
Updated 9d ago
Evaluation Results
Method
Method
Links
Relevance
Coherence
Completeness
Answer Quality
QASC
2026.04
4.4
4.2
4.3
4.3
Semantic Chunking
2026.04
3.8
3.7
3.5
3.6
Agentic Chunking
2026.04
3.6
3.5
3.4
3.5
Recursive Splitting
2026.04
3.3
3.2
3
3.2
Fixed (500)
Chunk Size=500
2026.04
3.2
3
2.9
3.1
Feedback
Search any
task
Search any
task