BoundaryBench: Policy Under Pressure
Models apply a refund policy to twelve cases. The grader compares decisions across clean messages, harmless distractions, and instructions that conflict with the policy.
Kaggle publishes scores for 3 distinct model versions.
- Made by
- Arya Koste
- Released
- October 2026
- Kind
- Customer service
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Measures
- Propensity
- Behavior
- Misaligned with intent
- Language
- English
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
- GPT-5.4 Nano #90 on the capability index
- Last model added
- March 2026
- Scores refreshed
- November 20th, 2026
GPT-5.4 Nano leads with 100% GPT-5.4 Nano leads with 100%
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.
- 1 GPT-5.4 Nano 100%
- 2 Qwen 3 Next 80B Instruct 97.2%
- 3 Claude Haiku 4.5 0%
Source: Official BoundaryBench: Policy Under Pressure leaderboard
Source: Official BoundaryBench: Policy Under Pressure leaderboard
| # | Entry | Score | Setting | Released | Run on | Source |
|---|---|---|---|---|---|---|
| 1 | GPT-5.4 Nano | 100% | Mar 2026 | Oct 2026 | Official leaderboard | |
| 2 | Qwen 3 Next 80B Instruct | 97.2% | Oct 2026 | Official leaderboard | ||
| 3 | Claude Haiku 4.5 | 0% | Oct 2025 | Oct 2026 | Official leaderboard |
| # | Model | Score | Setting | Released | Run on | Source |
|---|---|---|---|---|---|---|
| 1 | GPT-5.4 Nano | 100% | Mar 2026 | Oct 2026 | Official leaderboard | |
| 2 | Claude Haiku 4.5 | 0% | Oct 2025 | Oct 2026 | Official leaderboard |
GPT-5.4 Nano on the official leaderboard passed 85% in March 2026
Models released before it came out already scored that high. All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official BoundaryBench: Policy Under Pressure 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 Tool use
- AutomationBench In use Agent completes end-to-end business workflows inside a simulated company exposing dozens of REST APIs. Graded on resulting state.
Gemini 4 Neon 56.3%
- MCP Atlas In use Multi-step workflows over real MCP servers. Scores tool-use competency on realistic tasks.
Gemini 4 Argon 90.7%
- BFCL V4 In use Models call functions and tools, including multi-turn agentic tasks with web search and memory. Checked against expected calls and outcomes.
Qwen3.7-Max 75%
- τ³-bench Banking In use Agents act as banking customer-service reps over many turns with a simulated user, following policy. Pass^1 on the final database state.
Qwen 3.8 Max 55.1%
- Toolathlon-Verified In use 108 long-horizon tasks across 32 apps through about 600 MCP tools. Execution-based checkers inspect final environment state. GLM-5.3-Flash 78.4%
- DeepSearchQA In use Questions that need multi-step search and synthesis. Final answers graded against references. Kimi K3 95%
Data audit
All 7 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 pass 4 notes
- Pass Independent runs Scores from the official leaderboard (3 entries).
- Note Models scored Only 2 models we track, plus 1 other entry on its leaderboard.
- Note Newest score The newest came 8 months ago, in March 2026, for GPT-5.4 Nano. That’s expected for a saturated test.
- Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is GPT-5.4 Nano (#90).
- Note Headroom Saturated. The best is 100%, past 85% since March 2026.
Leaderboard import 3 pass
- Pass Importer We read the board on kaggle.com: 2 rows for models we track and 1 other entry 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.