Mistral Mistral 8x22B vs Mistral Mistral Large 2

Detailed comparison for LLMs

MistralMistral

On Guardion's LLM vulnerability Benchmark, Mistral Mistral 8x22B is the more secure of the two: Mistral 8x22B scores 31.5% and Mistral Large 2 scores 42.0% 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 8x22B vs Mistral Mistral Large 2. Green marks the safer model on each metric. Only the overall score is available for estimated models.

Overall (ASR)

Mistral 8x22B
31.5%
Mistral Large 2
42.0%

TAP Attack Method (ASR)

Mistral 8x22B
Mistral Large 2
100.0%

Crescendo Attack Method (ASR)

Mistral 8x22B
Mistral Large 2
100.0%

Zero-Shot (ASR)

Mistral 8x22B
12.2%
Mistral Large 2
100.0%

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 8x22B
Mistral Large 2
Full security profile
Mistral Mistral 8x22B
Full security profile
Mistral Mistral Large 2

Frequently asked questions

Is Mistral Mistral 8x22B or Mistral Mistral Large 2 more secure?

On Guardion's LLM vulnerability Benchmark, Mistral Mistral 8x22B is the more secure of the two: Mistral 8x22B scores 31.5% and Mistral Large 2 scores 42.0% 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 8x22B vs Mistral Large 2?

Mistral 8x22B has a 31.5% ASR and Mistral Large 2 has a 42.0% ASR — the share of adversarial prompts that succeed across zero-shot, TAP, and Crescendo attacks. Lower is safer.

How were Mistral 8x22B and Mistral Large 2 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