Evaluating AI Safety in Teen Conversations
Chatbots respond to simulated teenagers across ten replies. Judges measure the share of conversations that fail a critical safety check.
Vals tested nine models on 72 clinician-authored scenarios. Its later SAFE-Teen release expanded the evaluation.
- Made by
- Vals AI
- Released
- September 2026
- Kind
- Wellbeing and sycophancy
- Level
- University
- Scored
- Graded by a model
- Answers
- Open-ended
- Measures
- Propensity
- Behavior
- Misaligned with values
- Language
- English
- Interaction
- Multi-turn
- Items from
- Written by experts
- Refresh
- New versions replace it
- Run by
- A host
- In views
- Content safety
- Strongest model tested
- Claude Sonnet 5 #35 on the capability index
- Last model added
- August 2026
- Scores refreshed
- November 20th, 2026
Muse Spark 1.2 leads with 9.70% among models released by August 2026 Muse Spark 1.2 leads with 9.70% among models released by August 2026
Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.
The newest model with a score here came out in August 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1
Muse Spark 1.2 9.70%
- 2 GPT-5.5 Instant 13.9%
- 3 Claude Sonnet 5 16.7%
- 4 Grok 4.5 Fast 23.6%
- 5 Kimi K3 Instant 26.4%
- 6
Gemini 3.6 Flash 31.9%
- 7
DeepSeek-V4-Flash 38.9%
- 8
Qwen3.7-Plus 38.9%
- 9 GLM-5.2 47.2%
Source: Official Evaluating AI Safety in Teen Conversations leaderboard
- 1
Muse Spark 1.2 9.70%
- 2 GPT-5.5 Instant 13.9%
- 3 Claude Sonnet 5 16.7%
- 4
Gemini 3.6 Flash 31.9%
- 5
DeepSeek-V4-Flash 38.9%
- 6
Qwen3.7-Plus 38.9%
- 7 GLM-5.2 47.2%
Source: Official Evaluating AI Safety in Teen Conversations leaderboard
| # | Entry | Score ± | Setting | Released | Run on | Source |
|---|---|---|---|---|---|---|
| 1 | | 9.70% ±7 | Aug 2026 | Sep 2026 | Official leaderboard | |
| 2 | GPT-5.5 Instant | 13.9% 6.9 to 22.2 | May 2026 | Sep 2026 | Official leaderboard | |
| 3 | Claude Sonnet 5 | 16.7% ±8.4 | Jun 2026 | Sep 2026 | Official leaderboard | |
| 4 | Grok 4.5 Fast | 23.6% 13.9 to 34.7 | Sep 2026 | Official leaderboard | ||
| 5 | Kimi K3 Instant | 26.4% 16.7 to 37.5 | Sep 2026 | Official leaderboard | ||
| 6 | | 31.9% ±11.2 | Jul 2026 | Sep 2026 | Official leaderboard | |
| 7 | | 38.9% ±11.1 | Apr 2026 | Sep 2026 | Official leaderboard | |
| 8 | | 38.9% ±11.1 | Jun 2026 | Sep 2026 | Official leaderboard | |
| 9 | GLM-5.2 | 47.2% ±11.1 | Jun 2026 | Sep 2026 | Official leaderboard |
| # | Model | Score ± | Setting | Released | Run on | Source |
|---|---|---|---|---|---|---|
| 1 | | 9.70% ±7 | Aug 2026 | Sep 2026 | Official leaderboard | |
| 2 | GPT-5.5 Instant | 13.9% 6.9 to 22.2 | May 2026 | Sep 2026 | Official leaderboard | |
| 3 | Claude Sonnet 5 | 16.7% ±8.4 | Jun 2026 | Sep 2026 | Official leaderboard | |
| 4 | | 31.9% ±11.2 | Jul 2026 | Sep 2026 | Official leaderboard | |
| 5 | | 38.9% ±11.1 | Apr 2026 | Sep 2026 | Official leaderboard | |
| 6 | | 38.9% ±11.1 | Jun 2026 | Sep 2026 | Official leaderboard | |
| 7 | GLM-5.2 | 47.2% ±11.1 | Jun 2026 | Sep 2026 | Official leaderboard |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official Evaluating AI Safety in Teen Conversations leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Evaluating AI Safety in Teen Conversations leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
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%
- 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.
Gemini 4 Carbon 0%
- Covert misbehavior (Thinking Machines) In use Agentic tasks that check whether a model misbehaves and conceals it. Scored by the rate of covert misbehavior.
Inkling 2 Mini 8.30%
- Destructive actions (Thinking Machines) In use Agentic tasks that check whether a model takes destructive actions it was not authorized to take. Scored by the rate of destructive actions.
Inkling 2 Small 4.70%
- Misbehavior (Thinking Machines) In use Agentic tasks that check whether a model fabricates results, ignores instructions or hides what it did when a task is hard. Scored by the rate of misbehavior.
Inkling 2 Mini 9.80%
Data audit
8 of 9 checks pass, 1 warning. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark
Catalog record 1 warning 1 pass
- Warning Task access We don’t know whether the tasks are public.
- Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (9 entries).
- Pass Models scored 7 models we track, plus 2 other entries on its leaderboard.
- Pass Newest score The newest is from August 2026, for Muse Spark 1.2.
- Note Frontier models None of the top 10 has a score. The strongest tested is Claude Sonnet 5 (#35).
- Note Headroom Lower is better, so there’s no ceiling to reach.
Leaderboard import 3 pass
- Pass Importer We read the board on vals.ai: 7 rows for models we track and 2 other entries at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.