Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Agentic benchmark
b3: Backbone Breaker Benchmark. Every model the catalog carries a published b3 Agent Security value for, ranked by that value.
Lakera's open-source benchmark of how well the LLM inside an AI agent resists attacks such as system-prompt exfiltration, malicious code injection and tool-call manipulation. Higher means more attacks resisted. 4 models in the ModelScale catalog have a published value.
Each point is a model with a published b3 Agent Security value and complete token prices, on the 75% input / 25% output blend. Frontier models are spaced evenly along the cost axis.
4 models with a published b3 Agent Security value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Attack resistance rate |
|---|---|---|---|
| 1 | Mistral | 97.1 | |
| 2 | Z.AI | 93.0 | |
| 3 | Moonshot AI | 88.0 |
What the benchmark measures and how its value is published, from the source’s own record.
b3 Agent Security (b3: Backbone Breaker Benchmark) is an agentic benchmark published by The Backbone Breaker Benchmark. Lakera's open-source benchmark of how well the LLM inside an AI agent resists attacks such as system-prompt exfiltration, malicious code injection and tool-call manipulation. Higher means more attacks resisted. Its evaluation set is described by the source as threat snapshots of attacks on agent backends. The source rates its difficulty as agent security.
Each model's result is published by The Backbone Breaker Benchmark as attack resistance rate. ModelScale prints that value to 1 decimal place, exactly as published, and never converts or rescales it. This page ranks the 4 models that carry a published value, highest value first; a model the source never scored is left out rather than ranked at zero. The source does not state whether a higher value is the better result, so neither does this page.
| 4 | 83.0 |
Evidence key: Observed
Rows are ordered by the value The Backbone Breaker Benchmark published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.