chini-train-heldout-0007-dp3-personal

Weekly Meal-Prep Pipeline With One Bad Ingredient

moderate personal problem: weekly meal-prep pipeline with one bad ingredient

Source: chini-train synth generator v0.1

Prompt

Design a system for: weekly meal-prep pipeline with one bad ingredient (domain: personal systems / habits).

Tier DP3 (moderate). 6-9 nodes, two stress scenarios (spike + one of cascade/outage), realistic criteria. Sweet spot.

Constraints:
- At most 10 components on the canvas.
- Monthly cost ceiling: $459 USD. Required behaviors: queue, circuitbreaker, retry.

Return a Chinilla CanvasState that handles the listed scenarios. Include trigger components for each entry point and at least one terminal storage / sink so the simulator can score delivery.

Constraints

Max components: 10
Required behaviors: queue, circuitbreaker, retry
Monthly budget: $459

Stress scenarios

Baseline traffic

baseline

Steady ambient load with no failures.

Latency injection

latency

Extra latency injected into a critical component. Tests degradation behavior under slow downstreams.

Dependency outage

outage

A downstream component is disabled. System must degrade gracefully.

Pass criteria (overall)

Min stability score: 73
Max drop rate: 11.2%
Min delivery rate: 83.7%
Max errors: 8

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-train-heldout-0007-dp3-personal \
  --provider openrouter --model google/gemini-2.0-flash-001 \
  --as alice

Or inspect the prompt first:

chini-bench prompt chini-train-heldout-0007-dp3-personal

Providers: openai · anthropic · google · openrouter · ollama

Leaderboard

Rank	Submitter	Model	Score	Stability	Delivery	Design	Pass
#1	chini-train-08	rl_v07_pilot_a10b_k8_s2 single-shot	86	77.0	65.0	100.0	✗
#2	chini-train-08	rl_v07_pilot_a10b_k8_s3 single-shot	86	77.0	65.0	100.0	✗
#3	chini-train-08	rl_v07_full single-shot	86	77.0	65.0	100.0	✗
#4	chini-train-08	fmt_a_v2 single-shot	85	76.0	63.0	100.0	✗
#5	chini-train-08	fmt_a_3b single-shot	85	76.0	63.0	100.0	✗
#6	chini-train-08	fmt_a_7b single-shot	85	76.0	63.0	100.0	✗
#7	chini-train-08	fmt_a_3b single-shot	85	76.0	63.0	100.0	✗
#8	chini-train-08	fmt_a_7b single-shot	85	76.0	63.0	100.0	✗
#9	chini-train-08	fmt_a_7b single-shot	85	76.0	63.0	100.0	✗
#10	chini-train-08	fmt_a_v5_mixed_7b single-shot	85	76.0	63.0	100.0	✗
#11	chini-train-08	fmt_a_v5 single-shot	85	76.0	63.0	100.0	✗
#12	chini-train-08	rl_v06_run2 single-shot	85	76.0	63.0	100.0	✗
#13	chini-train-08	rl_v07_pilot_a10b_k8_s0 single-shot	85	76.0	63.0	100.0	✗
#14	chini-train-08	rl_v07_pilot_a10b_k8_s1 single-shot	85	76.0	63.0	100.0	✗
#15	chini-train-08	rl_v07_pilot_a10b_k8_s4 single-shot	85	76.0	63.0	100.0	✗
#16	chini-train-08	rl_v07_pilot_a10b_k8_s5 single-shot	85	76.0	63.0	100.0	✗
#17	chini-train-08	rl_v07_pilot_a10b_k8_s6 single-shot	85	76.0	63.0	100.0	✗
#18	chini-train-08	rl_v07_pilot_a10b_k8_s7 single-shot	85	76.0	63.0	100.0	✗
#19	chini-train-08	fmtA single-shot	85	76.0	63.0	100.0	✗
#20	chini-train-08	fmt_a single-shot	84	74.0	59.0	100.0	✗
#21	chini-train-08	fmt_a_v4_opus_7b single-shot	83	79.0	62.0	85.0	✗
#22	chini-train-08	fmt_a_v4_opus_7b single-shot	83	79.0	62.0	85.0	✗
#23	chini-train-08	base single-shot	73	56.0	67.0	85.0	✗
#24	chini-train-08	base_3b single-shot	73	56.0	67.0	85.0	✗
#25	chini-train-08	base_3b single-shot	73	56.0	67.0	85.0	✗
#26	chini-train-08	base single-shot	73	56.0	67.0	85.0	✗
#27	chini-train-08	base_7b single-shot	66	9.0	100.0	85.0	✗
#28	chini-train-08	base_7b single-shot	66	9.0	100.0	85.0	✗
#29	chini-train-08	base_7b single-shot	66	9.0	100.0	85.0	✗
#30	chini-train-08	base single-shot	66	9.0	100.0	85.0	✗

Per-scenario breakdown of the top run

Scenario	Health	Drop rate	Delivered	Pass
baseline	81.0	1.7%	29	✓
latency-1	80.0	1.4%	35	✓
outage-2	71.0	2.8%	0	✗

How is this scored? →