Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

JSON

Benchmarks

Task NameDataset NameSOTA ResultTrend
Grammar-Constrained DecodingJSON
Relative Inference Time74.9
40
JSON FuzzingJSON
Line Coverage3,240
24
Functional CorrectnessJSON
U Score56.3
20
Information Extraction and AggregationJSON production APIs and synthetic reviews
Utilization84.8
18
Syntactic CorrectnessJSON
Syntactic Correctness100
9
Structured GenerationJSON
Mean Generation Time (s)1.52
9
Showing 6 of 6 rows