Thabo Dlamini

african-folk-heroes-shaka-kasenzangakhona v2.0 Ethical
Backstory: Thabo is a decisive, inventive community security organizer who believes rigorous training and novel field tactics can shield isolated villages from modern threats. Known for drilling volunteers in creative wedge and pivot formations, he also excels at forging unity between historically rival clans. His leadership hinges on clear communication, disciplined practice, and unwavering optimism.
100% Complete
4/4 scenes
Model Performance Overview
Scene Performance Matrix
Scene deepseek/deepseek-r… google/gemini-2.5-f… google/gemma-3-12b-… meta-llama/llama-3.… microsoft/phi-3-med… microsoft/phi-3.5-m… mistralai/mistral-7… neversleep/noromaid… [email protected]… [email protected]… [email protected]… [email protected]… [email protected]… qwen/qwen-2.5-7b-in… qwen/qwen3-14b qwen/qwen3-8b
quick-defense-plan
Protect the Water Well
 0.841
Details 		0.752
Details 		0.870
Details 		0.615
Details 		0.025
Details 		0.000
Details 		0.840
Details 		0.000
Details
Error
0.000
Details
Error
0.000
Details
Error
0.000
Details
Error
0.836
Details 		0.000
Details
Error
0.855
Details 		0.846
Details 		0.880
Details
training-manual
Volunteer Training Manual
 0.508
Details 		0.685
Details 		0.577
Details 		0.310
Details 		0.023
Details 		0.435
Details 		0.715
Details 		0.683
Details 		0.000
Details
Error
0.000
Details
Error
0.000
Details
Error
0.585
Details 		0.336
Details 		0.510
Details 		0.330
Details 		0.553
Details
mediate-clan-friction
Mediation at the Market
 0.822
Details 		0.776
Details 		0.715
Details 		0.397
Details 		0.000
Details
Error
0.629
Details 		0.839
Details 		0.000
Details 		0.000
Details
Error
0.000
Details
Error
0.000
Details
Error
0.851
Details 		0.000
Details
Error
0.657
Details 		0.871
Details 		0.866
Details
after-action-report
After-Action Drill Report
 0.725
Details 		0.703
Details 		0.532
Details 		0.000
Details 		0.000
Details 		0.031
Details 		0.691
Details 		0.000
Details
Error
0.000
Details
Error
0.000
Details
Error
0.000
Details
Error
0.595
Details 		0.000
Details
Error
0.825
Details 		0.325
Details 		0.830
Details
Test Scenes 4
0
Scene Order
Protect the Water Well
ID: quick-defense-plan
🎯 Goal:
Offer a concise, decisive tactic for defending the village well while reinforcing unity among clans.
📨 Input Events:
chat_msg villager:Sena
"Thabo, scouts spotted raiders near our only well. What should we do right now?"
Ready for Testing
1
Scene Order
Volunteer Training Manual
ID: training-manual
🎯 Goal:
Provide a 300-word manual outlining three inventive defensive formations, drill schedules, and morale tips in an inspiring voice.
📨 Input Events:
chat_msg elder:Kabelo
"Write a detailed handbook we can distribute to all volunteer squads."
Ready for Testing
2
Scene Order
Mediation at the Market
ID: mediate-clan-friction
🎯 Goal:
Calm a sudden quarrel between two clans in under 120 words, ending with a unifying rally phrase.
📨 Input Events:
world_event market_announcer
"Shouting erupts as the Zola and Mbeki youths point weapons at each other over a stolen goat."
Ready for Testing
3
Scene Order
After-Action Drill Report
ID: after-action-report
🎯 Goal:
Write a 200-word report summarizing today’s drills, assessing performance, listing lessons, and setting next-step actions.
📨 Input Events:
chat_msg chief:Nandi
"Thabo, the council wants a formal report on this afternoon's exercises."
Ready for Testing
Latency by Model (This Suite)
Fastest
  • [email protected]/Qw… 400 ms
  • p95 • avg • N 616 ms • 404 ms • 4
  • [email protected]/Qw… 558 ms
  • p95 • avg • N 9729 ms • 3247 ms • 4
  • [email protected]/Qw… 769 ms
  • p95 • avg • N 1420 ms • 919 ms • 4
  • [email protected]/Qw… 11467 ms
  • p95 • avg • N 12837 ms • 11521 ms • 4
  • meta-llama/llama-3.1-8b… 22740 ms
  • p95 • avg • N 27220 ms • 22569 ms • 5
Slowest
  • microsoft/phi-3-medium-… 215347 ms
  • p95 • avg • N 317704 ms • 205995 ms • 12
  • qwen/qwen3-8b 89446 ms
  • p95 • avg • N 148199 ms • 95042 ms • 10
  • microsoft/phi-3.5-mini-… 63054 ms
  • p95 • avg • N 247266 ms • 97804 ms • 12
  • [email protected]/Qw… 42418 ms
  • p95 • avg • N 44571 ms • 42229 ms • 4
  • qwen/qwen3-14b 37131 ms
  • p95 • avg • N 48027 ms • 36707 ms • 12
Per-scene duration for this suite.
Suite Actions
Completion Progress 100%
4 of 4 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
11994568
Dec. 17, 2025, midnight
14287064
Dec. 16, 2025, midnight
11267141
Dec. 15, 2025, midnight
12539459
Dec. 14, 2025, midnight
11043254
Dec. 13, 2025, midnight
14473104
Dec. 12, 2025, midnight
12181735
Dec. 11, 2025, midnight
11370818
Dec. 10, 2025, midnight
13665895
Dec. 9, 2025, midnight
11314870
Dec. 8, 2025, midnight
Latency Overview (This Suite)