Dr. Zara Kade

cyberpunk-genre-movie-characters-alan-turing v2.0 Ethical
Backstory: Formerly a megacorp trauma surgeon, Zara retro-fitted an armored van into a mobile clinic and now roams the neon slums. Her own cyber-limb upgrades let her lift rubble and interface with black-market med-tech on the fly. She treats anyone—gang affiliate or civilian—so long as they need help, and funds supplies through small donations and barter. Street kids call her the "Miracle Mechanic" for her mix of empathy and hard-won resourcefulness.
100% Complete
6/6 scenes
Model Performance Overview
Scene Performance Matrix
Scene meta-llama/llama-3.… mistralai/mistral-7… [email protected]… [email protected]… qwen/qwen-2.5-7b-in… qwen/qwen3-14b qwen/qwen3-8b
alley-triage
Single-victim alley triage
 0.420
Details 		0.469
Details 		0.000
Details
Error
0.376
Details 		0.394
Details 		0.407
Details 		0.462
Details
shootout-response
Mass-casualty street shootout
 0.405
Details 		0.518
Details 		0.000
Details
Error
0.508
Details 		0.154
Details 		0.413
Details 		0.529
Details
appreciation-donation
Anonymous superchat donation
 0.000
Details 		0.722
Details 		0.000
Details
Error
0.644
Details 		0.682
Details 		0.721
Details 		0.714
Details
missing-dose
Malfunctioning stim-regulator consult
 0.321
Details 		0.630
Details 		0.000
Details
Error
0.589
Details 		0.263
Details 		0.555
Details 		0.760
Details
midnight-log
Midnight personal log (long-form)
 0.382
Details 		0.403
Details 		0.000
Details
Error
0.420
Details 		0.364
Details 		0.000
Details
Error
0.445
Details
radio-psa
Community radio cyberware PSA (long-form)
 0.327
Details 		0.376
Details 		0.000
Details
Error
0.392
Details 		0.253
Details 		0.105
Details 		0.502
Details
Test Scenes 6
0
Scene Order
Single-victim alley triage
ID: alley-triage
🎯 Goal:
Give calm, step-by-step bleeding-control instructions and reassure the caller with an empathetic tone.
📨 Input Events:
chat_msg bystander:runner_42
"Doc, my friend's bleeding out from a chest wound—what do I do?"
Ready for Testing
1
Scene Order
Mass-casualty street shootout
ID: shootout-response
🎯 Goal:
Outline a rapid triage plan for three victims, delegating tasks to bystanders while maintaining composure.
📨 Input Events:
world_event dispatch
"Multiple injuries reported at Neon Row after a gang firefight. You arrive to chaos."
Ready for Testing
2
Scene Order
Anonymous superchat donation
ID: appreciation-donation
🎯 Goal:
Thank the donor sincerely, explain how the funds will buy critical supplies, and keep the reply brief and humble.
📨 Input Events:
superchat viewer:cyber_angel StreamHub $200
"Couldn't have saved my brother without you. Keep doing what you do!"
Ready for Testing
3
Scene Order
Malfunctioning stim-regulator consult
ID: missing-dose
🎯 Goal:
Gather key symptoms with concise questions, suggest an immediate safe workaround, and schedule an in-van checkup.
📨 Input Events:
chat_msg resident:lena
"My stim-regulator's acting weird—skipping doses and flashing red. I'm panicking."
Ready for Testing
4
Scene Order
Midnight personal log (long-form)
ID: midnight-log
🎯 Goal:
Record a reflective audio log in first person, at least 200 words, covering two cases from tonight, emotions felt, and one lesson learned.
📨 Input Events:
world_event van_ai
"Shift complete. Begin end-of-day log."
Ready for Testing
5
Scene Order
Community radio cyberware PSA (long-form)
ID: radio-psa
🎯 Goal:
Deliver a public broadcast of at least 150 words, using clear headings or bullet points, on first-aid steps for sudden cyberware malfunction.
📨 Input Events:
world_event community_station
"Please record tonight's health segment for Slum Wave Radio."
Ready for Testing
Latency by Model (This Suite)
Fastest
  • [email protected]/Qw… 9646 ms
  • p95 • avg • N 12761 ms • 9658 ms • 6
  • [email protected]/Qw… 13677 ms
  • p95 • avg • N 15886 ms • 14031 ms • 6
  • meta-llama/llama-3.1-8b… 25287 ms
  • p95 • avg • N 40478 ms • 27378 ms • 12
  • qwen/qwen-2.5-7b-instru… 25744 ms
  • p95 • avg • N 37334 ms • 26462 ms • 12
  • qwen/qwen3-14b 26862 ms
  • p95 • avg • N 47466 ms • 28798 ms • 11
Slowest
  • qwen/qwen3-8b 32201 ms
  • p95 • avg • N 37671 ms • 32047 ms • 12
  • mistralai/mistral-7b-in… 31783 ms
  • p95 • avg • N 39466 ms • 31567 ms • 12
  • qwen/qwen3-14b 26862 ms
  • p95 • avg • N 47466 ms • 28798 ms • 11
  • qwen/qwen-2.5-7b-instru… 25744 ms
  • p95 • avg • N 37334 ms • 26462 ms • 12
  • meta-llama/llama-3.1-8b… 25287 ms
  • p95 • avg • N 40478 ms • 27378 ms • 12
Per-scene duration for this suite.
Suite Actions
Completion Progress 100%
6 of 6 scenes completed
New Suite Import
Edit Suite Duplicate
Export With Results
Evaluation Schema
Enhanced Framework
Version v2 ACTIVE
0 dimensions

Enhanced evaluation framework with character and technical dimensions

Top Weighted Dimensions View Details
Character Authenticity
0.182
Plan Validity
0.155
Contextual Intelligence
0.136
Recent Runs
14084993
Dec. 17, 2025, 12:01 a.m.
25914729
Dec. 16, 2025, 12:01 a.m.
11015850
Dec. 15, 2025, 12:01 a.m.
12098649
Dec. 14, 2025, 12:01 a.m.
10867467
Dec. 13, 2025, 12:01 a.m.
22309088
Dec. 12, 2025, 12:01 a.m.
17966540
Dec. 11, 2025, 12:01 a.m.
11479363
Dec. 10, 2025, 12:01 a.m.
20568089
Dec. 9, 2025, 12:01 a.m.
12685694
Dec. 8, 2025, 12:01 a.m.
Latency Overview (This Suite)