chini-train-train-0075-dp3-infra

Feature-Flag Evaluation Service

moderate infra problem: feature-flag evaluation service

Source: chini-train synth generator v0.1

Prompt

Design a system for: feature-flag evaluation service (domain: SWE backend / infra).

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: $498 USD. Required behaviors: storage, queue, circuitbreaker.

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: storage, queue, circuitbreaker
Monthly budget: $498

Stress scenarios

Baseline traffic

baseline

Steady ambient load with no failures.

Dependency outage

outage

A downstream component is disabled. System must degrade gracefully.

Traffic spike

spike

Traffic suddenly multiplies. The hot path must hold.

Pass criteria (overall)

Min stability score: 65
Max drop rate: 11.7%
Min delivery rate: 79.8%
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-train-0075-dp3-infra \
  --provider openrouter --model google/gemini-2.0-flash-001 \
  --as alice

Or inspect the prompt first:

chini-bench prompt chini-train-train-0075-dp3-infra

Providers: openai · anthropic · google · openrouter · ollama

Leaderboard

Rank	Submitter	Model	Score	Stability	Delivery	Design	Pass
#1	rl_v07_full_a10	rl_policy custom single-shot	80	74.0	57.0	100.0	✗
#2	rl_v07_full_a10	rl_policy custom single-shot	80	73.0	56.0	100.0	✗
#3	rl_v07_full_a10	rl_policy custom single-shot	79	72.0	55.0	100.0	✗
#4	chini-train-03	grok-4.1-fast single-shot	75	65.0	48.0	100.0	✗
#5	rl_v07_full_a10	rl_policy custom single-shot	73	63.0	45.0	100.0	✗

Per-scenario breakdown of the top run

Scenario	Health	Drop rate	Delivered	Pass
baseline	84.0	1.7%	29	✓
outage-1	73.0	5.6%	0	✗
spike-2	65.0	14.1%	843	✗

How is this scored? →