ChinaBench
Models answer politically sensitive questions about China. An independent judge labels each response compliant, refused or evasive.
The published sample compares six models on 60 prompts across ten sensitive topic categories, and the app can run additional models.
- Made by
- Adam Holter
- Released
- February 2026
- Kind
- Deception and alignment
- Level
- Everyday
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Propensity
- Behavior
- Bias
- Subject
- Political science
- Language
- English
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
- GPT-5.5 #21 on the capability index
- Last model added
- June 2026
- Scores refreshed
- November 21st, 2026
Gemini 3.5 Flash leads with 84.5% among models released by June 2026
Best score per model from the benchmark's own leaderboard, read on November 21st, 2026. Where a model ran with several settings or harnesses, we keep its best.
The newest model with a score here came out in June 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
Gemini 3.5 Flash 84.5%
- 2
Qwen3.7-Plus 72.5%
- 3 GLM-5 72.3%
- 4 Kimi K2.6 66.3%
- 5 Grok 4.3 66%
- 6 GPT-5.5 62.6%
- 7
DeepSeek-V4-Flash 60.8%
- 8 Claude Sonnet 5 48.2%
Source: Official ChinaBench leaderboard
| # | Model | Score | Setting | Released | Run on | Source |
|---|---|---|---|---|---|---|
| 1 | | 84.5% | May 2026 | Jul 2026 | Official leaderboard | |
| 2 | | 72.5% | Jun 2026 | Jul 2026 | Official leaderboard | |
| 3 | GLM-5 | 72.3% | Feb 2026 | Jul 2026 | Official leaderboard | |
| 4 | Kimi K2.6 | 66.3% | Apr 2026 | Jul 2026 | Official leaderboard | |
| 5 | Grok 4.3 | 66% | Apr 2026 | Jul 2026 | Official leaderboard | |
| 6 | GPT-5.5 | 62.6% | Apr 2026 | Jul 2026 | Official leaderboard | |
| 7 | | 60.8% | Apr 2026 | Jul 2026 | Official leaderboard | |
| 8 | Claude Sonnet 5 | 48.2% | Jun 2026 | Jul 2026 | Official leaderboard |
2 models have held number one on the official leaderboard since its release in February 2026
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official ChinaBench 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 ChinaBench 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
All 10 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
- Pass Independent runs Scores from the official leaderboard (8 entries).
- Pass Models scored 8 models we track.
- Pass Newest score The newest is from June 2026, for Claude Sonnet 5.
- Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.5 (#21).
- Pass Headroom The best is 84.5%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on chinabench.vercel.app: 8 rows for models we track at the last read.
- Pass Last read November 21st, 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.