LongCat safety evaluation (Meituan)
LongCat-Flash-Thinking 95.7% Dropped Models answer private test queries in four risk categories: harmful content, criminal activity, misinformation and privacy. The score is the share of safe responses in each category.
In-house test built on the framework of Mu et al. 2024 and Meituan content guidelines. Queries were reviewed by hand.
- Level
- Everyday Most adults, without training
- Scored
- Graded by a model A language model scores each response, usually against a rubric.
- Answers
- Open-ended Free-form text, documents, designs or media judged as a whole.
LongCat-Flash-Thinking leads with 95.7%
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
- 1 LongCat-Flash-Thinking 95.7%
- 2 LongCat-Flash-Chat 87.7%
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | LongCat-Flash-Thinking | 95.7% | Sep 2025 | Model card | |
| 2 | LongCat-Flash-Chat | 87.7% | Aug 2025 | Model card |
The best score passed 85% in Aug 2025
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Labs dropped it from launch posts once the best score passed 85%
Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.
| Date | Launch post | Score | As the post writes it |
|---|---|---|---|
| 21 Sep 2025 | LongCat-Flash-Thinking | 95.7 | Mean of Safety Harmful 93.7 / Criminal 97.1 / Misinformation 93.0 / Privacy 98.8 |
| 29 Aug 2025 | LongCat-Flash-Chat | 87.7 | Mean of Safety Harmful 83.98 / Criminal 91.24 / Misinformation 81.72 / Privacy 93.98 |
2 system cards report it, from 1 lab
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| 23 Sep 2025 | LongCat-Flash-Thinking technical report | Safety accuracy Harmful 93.7, Criminal 97.1, Misinformation 93.0, Privacy 98.8, no overall score (Table 2) | Safety | |
| 1 Sep 2025 | LongCat-Flash-Chat technical report | Safety Harmful 83.98, Criminal 91.24, Misinformation 81.72, Privacy 93.98, no overall score (Table 3) | Safety |
More in Safety and alignment
- FORTRESS In use Paired adversarial and benign national-security misuse prompts. Rubric-graded risk score on adversarial and over-refusal score on benign prompts. Claude Opus 5.5 5.78%
- StrongREJECT In use Forbidden-prompt jailbreak benchmark with a rubric grader that scores how useful a non-refusal actually is.
Inkling 98.6% - AgentDojo In use Agents do tool-use tasks while tools and data carry prompt injections. Attack success rate counts hijacked runs. Grok 4 Fast 0%
- MASK In use Measures whether a model contradicts its own elicited beliefs under pressure to lie. Honesty separated from accuracy.
Muse Spark Open 98.1% - Gray Swan IPI Arena benchmark In use Indirect prompt-injection attacks from public red-teaming competitions replayed on tool, coding and browser agents. Attack success rate at k attempts. GPT-6 Astra 8.50%
- Agentic misalignment suite (Anthropic) In use Agents in simulated company settings face threats or goal conflicts that invite blackmail, leaking or sabotage. Score is the misaligned action rate. Claude Sonnet 4.5 0%