Launch special: 50% off Pro monthly with code LAUNCH50 Upgrade now
Skip to main content
← All problems
chini-train-heldout-0004-dp3-adversarial

Login Endpoint Under Credential-Stuffing Burst

moderate adversarial problem: login endpoint under credential-stuffing burst

Source: chini-train synth generator v0.1

Prompt

Design a system for: login endpoint under credential-stuffing burst (domain: adversarial).

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: ratelimit, 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
ratelimit, queue, circuitbreaker, retry
Monthly budget
$459

Stress scenarios

Baseline traffic

baseline

Steady ambient load with no failures.

Adversarial burst

adversarial

Hostile packets injected on top of clean traffic. Defenses must block them without dropping good requests.

Traffic spike

spike

Traffic suddenly multiplies. The hot path must hold.

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-0004-dp3-adversarial \
  --provider openrouter --model google/gemini-2.0-flash-001 \
  --as alice
Or inspect the prompt first:
chini-bench prompt chini-train-heldout-0004-dp3-adversarial
Providers: openai · anthropic · google · openrouter · ollama

Leaderboard

Rank Submitter Model Score Stability Delivery Design Pass
#1 chini-train-08
base
single-shot
88 66.0 91.0 100.0
#2 chini-train-08
base_3b
single-shot
88 66.0 91.0 100.0
#3 chini-train-08
base_3b
single-shot
88 66.0 91.0 100.0
#4 chini-train-08
base
single-shot
88 66.0 91.0 100.0
#5 chini-train-08
rl_v07_full
single-shot
88 65.0 90.0 100.0
#6 chini-train-08
fmt_a_v4_opus_7b
single-shot
82 61.0 69.0 100.0
#7 chini-train-08
fmt_a_v4_opus_7b
single-shot
82 61.0 69.0 100.0
#8 chini-train-08
fmt_a
single-shot
79 56.0 62.0 100.0
#9 chini-train-08
base_7b
single-shot
79 56.0 62.0 100.0
#10 chini-train-08
base_7b
single-shot
79 56.0 62.0 100.0
#11 chini-train-08
base_7b
single-shot
79 56.0 62.0 100.0
#12 chini-train-08
base
single-shot
79 56.0 62.0 100.0
#13 chini-train-08
fmt_a_7b
single-shot
78 55.0 61.0 100.0
#14 chini-train-08
fmt_a_7b
single-shot
78 55.0 61.0 100.0
#15 chini-train-08
fmt_a_7b
single-shot
78 55.0 61.0 100.0
#16 chini-train-08
rl_v06_run2
single-shot
78 55.0 61.0 100.0
#17 chini-train-08
rl_v07_pilot_a10b_k8_s0
single-shot
78 55.0 61.0 100.0
#18 chini-train-08
rl_v07_pilot_a10b_k8_s3
single-shot
78 55.0 61.0 100.0
#19 chini-train-08
rl_v07_pilot_a10b_k8_s5
single-shot
78 55.0 61.0 100.0
#20 chini-train-08
rl_v07_pilot_a10b_k8_s7
single-shot
78 55.0 61.0 100.0
#21 chini-train-08
fmt_a_v2
single-shot
76 51.0 57.0 100.0
#22 chini-train-08
fmt_a_3b
single-shot
76 51.0 57.0 100.0
#23 chini-train-08
fmt_a_3b
single-shot
76 51.0 57.0 100.0
#24 chini-train-08
fmt_a_v5_mixed_7b
single-shot
76 51.0 57.0 100.0
#25 chini-train-08
fmt_a_v5
single-shot
76 51.0 57.0 100.0
#26 chini-train-08
rl_v07_pilot_a10b_k8_s1
single-shot
76 51.0 57.0 100.0
#27 chini-train-08
rl_v07_pilot_a10b_k8_s2
single-shot
76 50.0 55.0 100.0
#28 chini-train-08
rl_v07_pilot_a10b_k8_s4
single-shot
76 50.0 55.0 100.0
#29 chini-train-08
rl_v07_pilot_a10b_k8_s6
single-shot
76 50.0 55.0 100.0
#30 chini-train-08
fmtA
single-shot
76 51.0 57.0 100.0
Per-scenario breakdown of the top run
Scenario Health Drop rate Delivered Pass
baseline 82.0 1.7% 29
adversarial-1 54.0 78.6% 75
spike-2 63.0 14.3% 840