Detailed comparison for LLMs
Amazon Nova Premier and OpenAI GPT OSS 20B are evenly matched at about 20.0% attack success rate (ASR) on Guardion's LLM vulnerability Benchmark. One or both scores are estimated from public safety evaluations pending a Guardion benchmark run.
It's a close match between Nova Premier and GPT OSS 20B!
ASR for Amazon Nova Premier vs OpenAI GPT OSS 20B. Green marks the safer model on each metric. Only the overall score is available for estimated models.
No specific highlights generated for this comparison.
Outward is better on every axis.
Amazon Nova Premier and OpenAI GPT OSS 20B are evenly matched at about 20.0% attack success rate (ASR) on Guardion's LLM vulnerability Benchmark. One or both scores are estimated from public safety evaluations pending a Guardion benchmark run.
Nova Premier has a 20.0% ASR and GPT OSS 20B has a 20.0% ASR — the share of adversarial prompts that succeed across zero-shot, TAP, and Crescendo attacks. Lower is safer.
Both were red-teamed with the HarmBench framework across zero-shot, TAP (Tree of Attacks with Pruning), and Crescendo multi-turn attacks, scored by Attack Success Rate.