List your company

ResearchFormerly WizWand

Error Attribution

Benchmarks

DatasetSOTA methodMetricTrendResultsLast updated

62Removal Rate

30

May 11, 2026

GSM8k Right Dense

64Removal Rate

30

May 11, 2026

GSM8k Mid Dense

75Removal Rate

30

May 11, 2026

GSM8k Left Dense

82Removal Rate

30

May 11, 2026

Qwen3-8B-Thinking + GRPO

8.1Pair µF1

30

May 19, 2026

Qwen2.5-14B-Instruct + Aegis-SFT

16.62Pair µF1

30

May 19, 2026