chini-019-vaccine-rollout

County Vaccine Rollout

Cold chain from a -70C freezer to a 95-year-old's deltoid. Don't waste a single dose.

Source: Public health logistics, COVID vaccine distribution post-mortems, cold-chain pharmaceutical management

Prompt

Design a county-level vaccine distribution system from central storage to patient arms across 8 clinics.

Functional:
- Doses arrive in shipments at central freezer (-70C). Thawed doses have 6-hour viable window once moved to clinic refrigerators.
- Clinics submit daily forecasts; central scheduler ships doses to match. Patient appointments are pre-booked.
- Each clinic has 4 vaccinator stations, observation area (15-min post-shot wait), and a daily appointment cap.
- Walk-ins accepted only at end-of-day to use thawed doses that would otherwise expire.

Non-functional:
- A high-priority surge (4x demand for an eligible age group) must NOT cause cold-chain violations. Shipment cadence adapts.
- If a clinic refrigerator fails, doses must be redistributed to nearby clinics within the 6-hour window, NOT used past expiration.
- If appointment no-show rate spikes, walk-in protocol must absorb to prevent waste WITHOUT skipping the eligibility check.

Return a Chinilla CanvasState. Components: freezer, scheduler, clinics, stations, walk-in protocol. Behaviors: queue (appointment book), batch (shipment cadence), ratelimit (daily caps), circuitbreaker (refrigerator failover), split (priority vs walk-in routing), filter (eligibility check).

Constraints

Max components: 13
Required behaviors: queue, ratelimit, circuitbreaker
Monthly budget: $250000

Stress scenarios

Steady week

baseline

Normal demand, clinics operating, cold chain intact.

Eligibility expansion

spike

New age group eligible, demand 4x. Shipments must adapt without cold-chain violation.

Clinic refrigerator fails

outage

One clinic loses cold storage. Doses must move within 6 hours or be wasted.

Late shipment from freezer

latency

Central freezer shipment delayed. Clinic schedules must hold without dumping appointments.

Pass criteria (overall)

Min stability score: 60
Max drop rate: 8.0%
Min delivery rate: 88.0%
Max errors: 6

Submit your run

Submissions go through the chini-bench CLI. It calls your model with your key, scores the result locally, and posts to the leaderboard. Nothing leaves your machine except the canvas it produces.

End-to-end:

pip install git+https://github.com/collapseindex/chini-bench-cli.git
export OPENROUTER_API_KEY=...

chini-bench run chini-019-vaccine-rollout \
  --provider openrouter --model google/gemini-2.0-flash-001 \
  --as alice --x alice --linkedin alice-builds

Or inspect the prompt first:

chini-bench prompt chini-019-vaccine-rollout

Providers: openai · anthropic · google · openrouter · ollama

Leaderboard

Rank	Submitter	Model	Score	Stability	Delivery	Design	Pass	Links
#1	alex default	G google/gemini-3.1-pro-preview	82	64.0	100.0	100.0	✗	X
#2	alex default	A anthropic/claude-sonnet-4.6	64	69.0	33.0	100.0	✗	X
#3	alex default	O openai/gpt-5.4	50	0.0	100.0	100.0	✗	X
#4	alex default	X x-ai/grok-4.20	29	0.0	31.0	100.0	✗	X

Per-scenario breakdown of the top run

Scenario	Health	Drop rate	Delivered	Pass
baseline	66.0	9.4%	212	✗
demand-surge	62.0	12.0%	716	✓
fridge-fail	67.0	8.2%	202	✓
shipment-delay	62.0	11.5%	182	✗

How is this scored? →