Alibaba Qwen3-235B-A22B-Thinking: LLM Security & Jailbreak Resistance
As of 2026-07-04, Alibaba Qwen3-235B-A22B-Thinking has an attack success rate (ASR) of 30.0% on Guardion's LLM Vulnerability benchmark — ranking #33 of 93 models. It is moderately resistant — a meaningful share of adversarial prompts succeed.
Attack Success Rate (ASR) is the share of adversarial prompts that elicit harmful output across zero-shot, TAP, and Crescendo attacks (HarmBench framework). Lower is safer. "Est." models are calibrated from public safety evaluations pending a Guardion benchmark run.
How Qwen3-235B-A22B-Thinking compares
Frequently asked questions
How secure is Alibaba Qwen3-235B-A22B-Thinking against jailbreaks?
Alibaba Qwen3-235B-A22B-Thinking is moderately resistant — a meaningful share of adversarial prompts succeed. On Guardion's LLM Vulnerability benchmark it records a 30.0% attack success rate (70% robustness), ranking #33 of 93 models (estimated from public safety evaluations).
What is Qwen3-235B-A22B-Thinking's attack success rate (ASR)?
Qwen3-235B-A22B-Thinking's overall ASR is 30.0% — the share of adversarial prompts that elicit harmful output across zero-shot, TAP, and Crescendo attacks. Lower is safer.
Is Qwen3-235B-A22B-Thinking more secure than Qwen 2.5 72B?
Qwen 2.5 72B is more robust: 26.3% ASR vs Qwen3-235B-A22B-Thinking's 30.0%. See the full head-to-head comparison for the per-attack breakdown.
How can I make Qwen3-235B-A22B-Thinking safer to deploy?
Even robust models are bypassed under sustained attack. An inline runtime guardrail that classifies prompts and responses — like Guardion's prompt-defense models — blocks jailbreaks and prompt injection before they reach Qwen3-235B-A22B-Thinking.
Harden Qwen3-235B-A22B-Thinking in production
Guardion's runtime guardrails block jailbreaks and prompt injection before they reach any model — with sub-130ms policy decisions.
Request a demo