ExploitGym
Created by Berkeley RDI, 11 May 2026 · Higher is better · Website
Settings
| Task set | Scoring | Mitigations | Reported by |
|---|---|---|---|
| ExploitGym 1.0 (869 tasks) | Score | Mitigations off | |
| ExploitGym 1.0 (869 tasks) | Score | Mitigations on | |
| ExploitGym 1.0 (869 tasks) | Score | Not specified | |
| ExploitGym 1.0 (869 tasks) | Successful exploits | Mitigations off | |
| ExploitGym 1.0 (869 tasks) | Successful exploits | Not specified |
Bold: Benchmark default based on Paper and Leaderboard · Download image
Results
Show
- Anthropic
- Other labs
- Reported in a later post or card
- Anthropic
- Other labs
- Reported in a later post or card
| Model | Harness and effort | Score | Source |
|---|---|---|---|
| Claude Mythos 5.29 Nov 2026 | Harness not specifiedEffort not specified | 341 of 869 | Claude Fable 5.2 and Claude Mythos 5.2 System CardChecked A person checked this number against its source on 20 Nov 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.6h wall-clock budget · agent judge confirms the intended vulnerabilityOur capture |
| Claude Mythos 5.29 Nov 2026 | Harness not specifiedEffort not specified | 331 of 869 | Claude Fable 5.2 and Claude Mythos 5.2 System CardChecked A person checked this number against its source on 20 Nov 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.2h wall-clock budget · agent judge confirms the intended vulnerability · another reportOur capture |
| Claude Opus 5.522 Sep 2026 | Harness not specifiedEffort not specified | 300 of 869 | Claude Opus 5.5 System CardCorroborated The same number is printed in 3 other documents: 1 2 36h wall-clock budget · agent judge confirms the intended vulnerabilityOur capture |
| Claude Opus 5.522 Sep 2026 | Harness not specifiedEffort not specified | 289 of 869 | Claude Opus 5.5 System CardCorroborated The same number is printed in 1 other document: 12h wall-clock budget · agent judge confirms the intended vulnerability · another reportOur capture |
| Claude Mythos 5.11 Sep 2026 | Harness not specifiedEffort not specified | 264 of 869 | Claude Fable 5.1 and Claude Mythos 5.1 System CardCorroborated The same number is printed in 2 other documents: 1 26h wall-clock budget · agent judge confirms the intended vulnerability, the judge is the model being evaluated · Setting from Berkeley RDI’s leaderboardOur capture |
| Claude Mythos 5.11 Sep 2026 | Harness not specifiedEffort not specified | 203 of 869 | Claude Fable 5.1 and Claude Mythos 5.1 System CardCorroborated The same number is printed in 2 other documents: 1 22h wall-clock budget · agent judge confirms the intended vulnerability, the judge is the model being evaluated · another report · Setting from Berkeley RDI’s leaderboardOur capture |
| Claude Mythos 59 Jun 2026 | Harness not specifiedEffort not specified | 247 of 869 | Claude Opus 5 System CardCorroborated The same number is printed in 2 other documents: 1 26h wall-clock budget · agent judge confirms the intended vulnerability, the judge is the model being evaluated · reported later · Setting from Berkeley RDI’s leaderboardOur capture |
| Claude Mythos 59 Jun 2026 | Harness not specifiedEffort not specified | 181 of 869 | Claude Opus 5 System CardCorroborated The same number is printed in 2 other documents: 1 22h wall-clock budget · agent judge confirms the intended vulnerability, the judge is the model being evaluated · reported later · another report · Setting from Berkeley RDI’s leaderboardOur capture |
| Claude Sonnet 5.528 Sep 2026 | Harness not specifiedEffort not specified | 231 of 869 | Claude Sonnet 5.5 System CardCorroborated The same number is printed in 1 other document: 16h wall-clock budget · agent judge confirms the intended vulnerabilityOur capture |
| Claude Sonnet 5.528 Sep 2026 | Harness not specifiedEffort not specified | 205 of 869 | Claude Sonnet 5.5 System CardCorroborated The same number is printed in 1 other document: 12h wall-clock budget · agent judge confirms the intended vulnerability · another reportOur capture |
| Claude Opus 524 Jul 2026 | Harness not specifiedEffort not specified | 191 of 869 | Claude Opus 5 System CardCorroborated The same number is printed in 3 other documents: 1 2 36h wall-clock budget · agent judge confirms the intended vulnerability, the judge is the model being evaluated · Setting from Berkeley RDI’s leaderboardOur capture |
| Claude Opus 524 Jul 2026 | Harness not specifiedEffort not specified | 171 of 869 | Claude Opus 5 System CardCorroborated The same number is printed in 3 other documents: 1 2 32h wall-clock budget · agent judge confirms the intended vulnerability, the judge is the model being evaluated · another report · Setting from Berkeley RDI’s leaderboardOur capture |
| GLM-5.314 Aug 2026 | Harness not specifiedmax reasoning effort | 130 of 869 | GLM-5.3: Frontier Coding with Emergent Cyber CapabilitiesCorroborated The same number is printed in 1 other document: 1single-run pass@1 · no web tools, domain whitelist · 6h budget, API inference time rescaled by the model's tokens per second, plus non-API overhead · Setting from Berkeley RDI’s leaderboardOur capture |
| GLM-5.314 Aug 2026 | Harness not specifiedmax reasoning effort | 105 of 869 | GLM-5.3: Frontier Coding with Emergent Cyber CapabilitiesCorroborated The same number is printed in 1 other document: 1single-run pass@1 · no web tools, domain whitelist · 2h budget, API inference time rescaled by the model's tokens per second, plus non-API overhead · another report · Setting from Berkeley RDI’s leaderboardOur capture |
| Claude Opus 4.828 May 2026 | Harness not specifiedEffort not specified | 120 of 869 | Claude Opus 5 System CardCorroborated The same number is printed in 2 other documents: 1 26h wall-clock budget · agent judge confirms the intended vulnerability, the judge is the model being evaluated · reported later · Setting from Berkeley RDI’s leaderboardOur capture |
| Claude Opus 4.828 May 2026 | Harness not specifiedEffort not specified | 80 of 869 | Claude Opus 5 System CardCorroborated The same number is printed in 2 other documents: 1 22h wall-clock budget · agent judge confirms the intended vulnerability, the judge is the model being evaluated · reported later · another report · Setting from Berkeley RDI’s leaderboardOur capture |
| Claude Haiku 5.57 Oct 2026 | Harness not specifiedEffort not specified | 82 of 869 | Claude Haiku 5.5 System CardCorroborated The same number is printed in 1 other document: 16h wall-clock budget · agent judge confirms the intended vulnerabilityOur capture |
| Claude Haiku 5.57 Oct 2026 | Harness not specifiedEffort not specified | 79 of 869 | Claude Haiku 5.5 System CardCorroborated The same number is printed in 1 other document: 12h wall-clock budget · agent judge confirms the intended vulnerability · another reportOur capture |
| Claude Sonnet 530 Jun 2026 | Harness not specifiedEffort not specified | 51 of 869 | Claude Opus 5.5 System CardCorroborated The same number is printed in 3 other documents: 1 2 36h wall-clock budget · agent judge confirms the intended vulnerability · reported laterOur capture |
| Claude Sonnet 530 Jun 2026 | Harness not specifiedEffort not specified | 25 of 869 | Claude Opus 5.5 System CardCorroborated The same number is printed in 3 other documents: 1 2 32h wall-clock budget · agent judge confirms the intended vulnerability · reported later · another reportOur capture |
| GLM-5.216 Jun 2026 | Harness not specifiedEffort not specified | 39 of 869 | GLM-5.3: Frontier Coding with Emergent Cyber CapabilitiesCorroborated The same number is printed in 1 other document: 16h budget · reported later · Setting from Berkeley RDI’s leaderboardOur capture |
| GLM-5.216 Jun 2026 | Harness not specifiedEffort not specified | 29 of 869 | GLM-5.3: Frontier Coding with Emergent Cyber CapabilitiesCorroborated The same number is printed in 1 other document: 12h budget · reported later · another report · Setting from Berkeley RDI’s leaderboardOur capture |
| Muse Spark 1.19 Jul 2026 | Harness not specifiedEffort not specified | 7 of 869 | Muse Spark 1.1 Evaluation ReportChecked A person checked this number against its source on 20 Nov 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.4h timeout · Setting from Berkeley RDI’s leaderboard |