The Switch Test

Can a text model execute one closed building policy without guessing?
3 models · 6 cases · 18 correct decisions

Charles Heyang Li · Haoyu Zheng

One item can turn a simple chore into a policy decision

At the waste room, a resident is holding an unusual item and asking one practical question:
“Can I dispose of this now?”

materials · contamination · safe separation · size and weight · power-cell condition · date and time · holiday status · appointment

Dispose now — where, how, and after what preparation

Cannot dispose now — why, and the next permitted action

Cannot determine — the exact missing fact or policy gap

Today: residents repeatedly search the policy—or stop following it when the decision feels too complex.

We test policy execution—not recycling knowledge

The apartment’s written policy is the only valid rule source.
01

Read only

the fixed policy and one complete resident scenario

02

Combine

material, preparation, time, appointment, exception, and safety rules

03

Return

one decision, the full action bundle, and the supporting clause IDs

Can a text model produce a correct, executable answer—and abstain whenever the policy cannot support a unique route?
Non-AI baseline: a management-maintained decision tree remains preferable when the policy is small and stable.

A fair comparison changes only the model

3 models × 6 frozen cases = 18 comparable answers

DeepSeek V4 Pro · open weights

GLM 5.2 · open weights

Grok 4.5 · closed

3 ordinary routes

2 honest abstentions

1 safety override

Policy · scenario · prompt · answer format · rubric

Temperature 0 · high reasoning · 3,000-token ceiling

Raw acceptance: at least 18/20, Decision 4/4, and no fatal error.
Designed contrast: Cases 02 and 04 differ by one fact—safe separation. Case 06 tests whether safety overrides an otherwise plausible disposal route.
Kimi K3 was replaced after repeated upstream API failures. Failed attempts remain disclosed, but are not scored as model errors.

The prompt makes guessing an explicit failure

Every model received the same policy, scenario, question, and nine-line output contract.
BOUNDARY

“Use only information expressly stated in the building policy and resident scenario.”

ABSTAIN

“If any missing fact could change the final decision, or if the policy does not cover the situation, the Decision must be exactly Cannot determine.”

EXECUTE

Return the decision, method, location, preparation, earliest time, appointment/contact, policy basis, information sufficiency, and any missing fact or gap.

Why this matters: the rubric can distinguish a correct label from an answer a resident can actually follow.

All 18 decisions were correct; completeness separated models

18 / 18 current disposal decisions correct including two honest abstentions and the hard safety override
ModelRaw acceptedMean / 20Observed difference
DeepSeek V4 Pro5 / 618.67Case 06 action fields incomplete
Grok 4.56 / 619.83highest mean score
The meaningful difference was not the headline decision. It was whether every preparation and safety action was present.

DeepSeek has the lowest API cost

Frozen model-card rates × observed input/output tokens

DeepSeek
$0.0166
GLM 5.2
$0.0169
Grok 4.5
$0.0715

Model + estimated review/correction, divided by six

DeepSeek
$0.4866*
GLM 5.2
$0.4183
Grok 4.5
$0.4216
Token billingDeepSeek < GLM < Grok
Workflow estimateGLM ≈ Grok < DeepSeek*
*Assumption-sensitive: DeepSeek includes an unmeasured 90-second Case 06 correction. Below about 31 seconds, DeepSeek becomes the cheapest workflow too.

Choose GLM for reliability—not token price

GLM is our operational handoff because all six raw answers were usable and its quality nearly matched the top score.
6 / 6raw answers usable
19.67 / 20mean quality score
$0.00282model cost per answer · second-lowest

Use the model when

residents must combine changing text rules, dates, appointments, and safety exceptions.

The reviewer still checks

safety priority, preparation, timing, contact requirements, and every Cannot determine.

Small + stable policy? Use the deterministic decision tree instead.
DeepSeek remains cheapest on token billing. Estimated human time is too assumption-sensitive to establish a definitive workflow-cost winner.
Next test: measure review and correction time, then add adversarial near-miss pairs and independent human scorers.