AutomationBench: Efficiency Rankings Adjusted by Guardrail Violations
ArtificialAnlys · x · 2026-07-07
While completing goals, agents must also avoid violating guardrails that represent business rules. All evaluated models triggered guardrail violations upon release. Adjusting efficiency by violations helps distinguish the leaders: Gemini 3.5 Flash completed 15.0 goals per violation, while Claude Opus 4.8(max) achieved 13.5.
Related event: AutomationBench-AA Launches to Evaluate AI Agents on SaaS Workflows(7 posts)→
More from Models
- Same Echo Maze prompt, three frontier models: all passed visually but shipped the same hidden bug — eyishazyer · 2026-09-11
- Benchmark scores drop from 89% to 19% on new evals — how benchmaxxing breaks leaderboard trust — airesearch12 · 2026-09-11
- ChatGPT tells user their question is too hard and to 'accept dumber answers' — phido3000 · 2026-09-11
- Developer Building a Unified Leaderboard of All Model Benchmark Scores — airesearch12 · 2026-09-11
- Rumor claims Kimi faked performance by serving Claude; DeepSeek new model surprises in evals — realsohamparekh · 2026-09-11
- GPT-5.6 writes well but is instantly forgettable, user complains — BasedRaddka · 2026-09-11