ExploitGym
AI benchmark results and model performance
Type: percentageSecurity
Recorded results
2
Average score
35.05%
Score range
15.30% - 54.80%
Leading model
54.80% - Gemini 4 Carbon
Scores over time
Individual benchmark scores plotted by date.
Model results
| Organisation | Models | Reported | Top score | Info | Self reported | Source |
|---|---|---|---|---|---|---|
| Gemini 4 Carbon | 28 Oct 2026 | 54.80% | Google-reported launch evaluation; Evaluated without guardrails; source chart: https://storage.googleapis.com/gweb-uniblog-publish-prod/images/gemini_4_carbon__evals-table.original.png; methodology: deepmind.google/models/evals-methodology/gemini-4-carbon | Yes | Source | |
| DeepSeek V4.1 Flash | 10 Sept 2026 | 15.30% | Instruct model; reasoning_effort=100; Pass@1 | Yes | Source |
