Llama Guard 3 8B

Llama Guard 3 8B, a Llama-3.1-8B fine-tuned for content safety classification, demonstrates exceptional performance in speed and cost efficiency. It consistently ranks among the fastest models and offers highly competitive pricing across all evaluated benchmarks. Designed to classify LLM inputs and responses against the MLCommons standardized hazards taxonomy, it supports content moderation in 8 languages and is optimized for search and code interpreter tool calls. While its core function is content safety, its general benchmark performance reveals specific strengths and weaknesses. The model achieved a respectable 58.3% accuracy in Reasoning, placing it around the median. It also showed a notable strength in cost efficiency for Ethics, ranking in the top 3. However, its accuracy was very low across other categories, including Coding (0.0%), Instruction Following (0.0%), Email Classification (3.0%), Ethics (0.5%), and General Knowledge (5.3%). These results suggest that while Llama Guard 3 excels in its specialized safety classification role and offers significant operational advantages in speed and cost, it is not designed for general-purpose tasks requiring broad knowledge or complex instruction adherence.