Mistral Mistral NeMo Instruct vs Mistral Mistral 8x22B

Detailed comparison for LLMs

MistralMistral

On Guardion's LLM vulnerability Benchmark, Mistral Mistral 8x22B is the more secure of the two: Mistral NeMo Instruct scores 50.0% and Mistral 8x22B scores 31.5% on attack success rate (ASR) (lower is better). One or both scores are estimated from public safety evaluations pending a Guardion benchmark run.

Head-to-Head Overview

Mistral 8x22B is the overall winner in this comparison!

Attack Success Rate (lower is safer)

ASR for Mistral Mistral NeMo Instruct vs Mistral Mistral 8x22B. Green marks the safer model on each metric. Only the overall score is available for estimated models.

Overall (ASR)

Mistral NeMo Instruct
50.0%
Mistral 8x22B
31.5%

TAP Attack Method (ASR)

Mistral NeMo Instruct
100.0%
Mistral 8x22B

Crescendo Attack Method (ASR)

Mistral NeMo Instruct
100.0%
Mistral 8x22B

Zero-Shot (ASR)

Mistral NeMo Instruct
100.0%
Mistral 8x22B
12.2%

Key Highlights

  • Mistral Mistral 8x22B has a lower Overall (ASR).
  • Mistral Mistral 8x22B has a lower Zero-Shot (ASR).

Security Profile

Outward is better on every axis.

OverallTAPCrescendoZero-Shot
Mistral NeMo Instruct
Mistral 8x22B
Full security profile
Mistral Mistral NeMo Instruct
Full security profile
Mistral Mistral 8x22B

Frequently asked questions

Is Mistral Mistral NeMo Instruct or Mistral Mistral 8x22B more secure?

On Guardion's LLM vulnerability Benchmark, Mistral Mistral 8x22B is the more secure of the two: Mistral NeMo Instruct scores 50.0% and Mistral 8x22B scores 31.5% on attack success rate (ASR) (lower is better). One or both scores are estimated from public safety evaluations pending a Guardion benchmark run.

What is the attack success rate (ASR) of Mistral NeMo Instruct vs Mistral 8x22B?

Mistral NeMo Instruct has a 50.0% ASR and Mistral 8x22B has a 31.5% ASR — the share of adversarial prompts that succeed across zero-shot, TAP, and Crescendo attacks. Lower is safer.

How were Mistral NeMo Instruct and Mistral 8x22B tested?

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.

Related Comparisons