CyberGym
Cybersecurity evaluation of vulnerability reproduction.
- Category
- Other
- Unit
- Percent (%)
- Higher is better
- Comparison groups
- 2total
- All model rankings (records) · Ranked results: 2
All model rankings (records)
These rankings are recalculated by comparison group including legacy models and models unavailable in Claude Code / Codex. They differ from the rankings above because the set of models differs.
Show rankings including legacy and unavailable models (entries: 2)
official results (GLM-5.3 (Z.ai docs))
| Rank | Model | effort | Score | Source cost (metric) | measured/reported by | source |
|---|---|---|---|---|---|---|
| 1 | GLM-5.3 | not stated | 84.5% | — | Z.ai | GLM-5.3 (Z.ai docs)Retrieved 2026-11-12 · Vendor-reported value. Self-reported by Z.ai. effort not stated (default is max)Source excerptCyberGym 84.5% |
official results (Introducing Claude Opus 4.7)
| Rank | Model | effort | Score | Source cost (metric) | measured/reported by | source |
|---|---|---|---|---|---|---|
| 1 | Claude Opus 4.6 | not stated | 73.8% | — | Anthropic | Introducing Claude Opus 4.7Retrieved 2026-11-13 · Vendor-reported value (announced by Anthropic). Text values in the main text and the footnote (the table and figure were not used because they are images). Value after updating harness parameters. Effort not stated.Source excerptCyberGym: Opus 4.6’s score has been updated from the originally reported 66.6 to 73.8, as we updated our harness parameters to better elicit cyber capability.Location Opus 4.7 announcement page, CyberGym entry in Footnotes |