The Research
The research behind ReGrade
Each result says which dataset it comes from and how many models it covers.
The research, page by page
Cleared for public release
Joint study with AFRL
ReGrade® helps AI coding agents fix 152% more bugs.
Measured on 17 models.
See the evidence →Curtail
More bugs fixed
In our ongoing testing,
ReGrade® helps AI coding agents fix 130% more bugs.
Measured on 84 of the 84 models tested so far.
See the evidence →Curtail
Cost reduction
In our ongoing testing,
ReGrade® cuts LLM API cost per bug fixed by over 58%.
Measured on 75 of the 84 models tested so far.
See the evidence →Curtail
AI coding hallucinations
In our ongoing testing,
ReGrade® reduces AI coding hallucinations by over 83%.
Measured on 84 of the 84 models tested so far.
See the evidence →Pages marked Ongoing testing are recomputed as we test more models. The published study’s figures do not change.
See what changed in your own traffic.
Choose one service and define a baseline, candidate build and replay traffic. Evaluate repair results and total operating cost before expanding to your team.
See what ReGrade finds in your traffic.