Karmendra Pandey
We track 3 benchmarks that Karmendra Pandey made or helped make.
0 in use 0 with public code 2 with a leaderboard here
Benchmarks
| Name | Kind | Released | Best score | Status |
|---|---|---|---|---|
| Argus-Bench v2 Models screen inputs to a language model gateway. The benchmark checks detection rates, false positives, and agreement with the policy on severity. | Jailbreaks and prompt injection | Nov 2026 | | Not in cards |
| Cost Per Correct v2 Models answer tasks at three difficulty levels and repeat some with paraphrased instructions. Rules check correctness so the author can compare cost per correct answer. | General assistants | Nov 2026 | $91 | Not in cards |
| Cost Per Correct Models complete five tasks with verifiable answers. The evaluation measures the price of obtaining a correct answer. | General assistants | Oct 2026 | Not in cards |