Comparing Large Models' Practical Cybersecurity Capabilities
daniel_mac8 · x · 2026-07-17
The tweet cites results from AISI's (AI Safety Institute) cybersecurity range test "The Last Ones." Data shows that GLM-5.2 matches the performance of Claude Opus 4.5 released about 7 months ago, while DeepSeek V4-Pro lags behind Claude Sonnet 4.5 from the same period. The author believes such safety tests are the most objective benchmarks for assessing models' actual destructive power in cyberspace.
Related event: AISI says open models narrow the cyber-range gap(6 posts)→
More from Models
- Astra Scores 83% on GauntletBench, First Computer-Use Agent to Beat Human Baseline — ducha_aiki · 2026-09-11
- Kimi K2.8 Preview rolls out: near-K3 coding performance, 1M context for all tiers — teortaxesTex · 2026-09-11
- Looking for a classifier of software engineering task shapes to pick models per task — StewartalsopIII · 2026-09-11
- DeepSeek V4 Pro API to continue after Sept 2026, billing unchanged — teortaxesTex · 2026-09-11
- DeepSeek V4.1 Flash Hits 98% of GPT-6 Astra's Score at 1.4% of the Cost in Third-Party Benchmark — ayushtweetshere · 2026-09-11
- TheZvi Polls: Has Your Coding Model Choice Changed Since Fable 5.1 and Astra? — TheZvi · 2026-09-11