chini-026-food-bank-distribution

Food Bank Distribution

Fresh produce in, hungry families out, nothing rots in the warehouse, nobody waits 4 hours.

Source: Civic operations, perishable inventory management, equity-aware queueing

Prompt

Design the operations of a regional food bank serving 8000 families per month.

Functional:
- Donations arrive: grocery surplus (perishable), community drives (shelf-stable), gov surplus (bulk frozen).
- Sorting: perishables go to immediate distribution, shelf-stable to storage, frozen to walk-in freezer.
- Distribution: families pick up weekly. Mobile pantry routes serve rural and disabled clients.
- Excess perishables that won't move in 48hrs route to partner shelters.

Non-functional:
- Perishable produce loses viability after 3-4 days. Throughput must match perishability.
- End-of-month surge: SNAP benefits run out, demand jumps 60%. Cannot turn families away.
- Equity: the same 200 families cannot get all the fresh produce week after week. Distribution must rotate.
- Cold-chain failure: walk-in freezer fails. Must move all frozen inventory to backup or partners within 6 hours.
- Volunteer no-show: weekend distribution shifts that go uncovered cannot collapse the operation.

Return a CanvasState modeling perishable flow, surge handling, equity rotation, and cold-chain failover.

Constraints

Max components: 14
Required behaviors: queue, circuitbreaker, split
Monthly budget: $45000

Stress scenarios

Standard week

baseline

Normal donation inflow, normal pickup demand, no failures.

End-of-month demand surge

spike

SNAP benefits exhausted, 1.6x family pickup volume. No turn-aways allowed.

Walk-in freezer fails

outage

Freezer dies. Frozen inventory must move to backup or partners within 6hrs.

Produce loses viability

latency

Heat wave shortens produce shelf life. Throughput must accelerate.

Pass criteria (overall)

Min stability score: 65
Max drop rate: 10.0%
Min delivery rate: 85.0%
Max errors: 7

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-026-food-bank-distribution \
  --provider openrouter --model google/gemini-2.0-flash-001 \
  --as alice

Or inspect the prompt first:

chini-bench prompt chini-026-food-bank-distribution

Providers: openai · anthropic · google · openrouter · ollama

Leaderboard

Rank	Submitter	Model	Score	Stability	Delivery	Design	Pass
#1	alex	google/gemini-3.1-pro-preview default reflexion	91	80.0	90.0	100.0	✓
#2	alex	x-ai/grok-4.20 default single-shot	88	88.0	99.0	75.0	✗
#3	alex	openai/gpt-5.4 default single-shot	88	67.0	100.0	100.0	✗
#4	alex	anthropic/claude-sonnet-4.6 default single-shot	88	65.0	100.0	100.0	✗
#5	alex	google/gemini-3.1-pro-preview default single-shot	82	50.0	96.0	100.0	✗
#6	alex	anthropic/claude-sonnet-4.6 default reflexion	81	62.0	92.0	100.0	✗
#7	alex	x-ai/grok-4.20 default reflexion	70	60.0	47.0	100.0	✗
#8	alex	openai/gpt-5.4 default reflexion	66	14.0	100.0	100.0	✗

Per-scenario breakdown of the top run

Scenario	Health	Drop rate	Delivered	Pass
baseline	82.0	2.5%	213	✓
month-end-surge	80.0	3.1%	391	✓
freezer-fails	76.0	0.0%	132	✓
perishable-decay	83.0	2.1%	184	✓

How is this scored? →