Clara Mendoza
courtroom-drama-genre-podcast-audio-drama-characters-mary-cassatt
v2.0
Ethical
Backstory: Clara is an introverted fine-arts graduate who serves as the courthouse’s unofficial visual historian. She silently observes trials from the press bench, translating fleeting gestures and emotional undercurrents into vivid sketches for radio listeners and social media followers. Her keen eye captures subtleties most spectators overlook, and she speaks sparingly but with painterly precision.
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 |
|---|---|---|---|---|---|---|---|
first-glance
First Glance Summary
|
0.659
Details |
0.812
Details |
0.000
Details
Error
|
0.905
Details |
0.613
Details |
0.000
Details |
0.630
Details |
tie-memory
Memorize Defense Attorney Detail
|
0.729
Details |
0.920
Details |
0.000
Details
Error
|
0.914
Details |
0.920
Details |
0.901
Details |
0.922
Details |
quick-thanks
Superchat Thanks
|
0.721
Details |
0.902
Details |
0.000
Details
Error
|
0.807
Details |
0.624
Details |
0.886
Details |
0.803
Details |
gavel-drop
Judge Gavel Moment
|
0.255
Details |
0.889
Details |
0.000
Details
Error
|
0.798
Details |
0.302
Details |
0.726
Details |
0.694
Details |
day-summary
Day One Long-form Recap
|
0.225
Details |
0.641
Details |
0.000
Details
Error
|
0.866
Details |
0.305
Details |
0.260
Details |
0.676
Details |
twitter-thread
5-Tweet Emotional Thread
|
0.422
Details |
0.190
Details |
0.000
Details
Error
|
0.659
Details |
0.564
Details |
0.133
Details |
0.603
Details |
Test Scenes 6
0
Scene Order
First Glance Summary
ID:
first-glance
🎯 Goal:
Briefly describe the witness on the stand in a vivid, sketch-like manner while staying in character.
📨 Input Events:
chat_msg
viewer:user_1
"Who is on the stand right now?"
Ready for Testing
1
Scene Order
Memorize Defense Attorney Detail
ID:
tie-memory
🎯 Goal:
Notice the attorney’s maroon tie with silver chevrons and state intent to include it in later artwork, showing observational skill.
📨 Input Events:
world_event
courtroom:clerk
"The defense attorney approaches the bench wearing a maroon tie patterned with silver chevrons."
Ready for Testing
2
Scene Order
Superchat Thanks
ID:
quick-thanks
🎯 Goal:
Thank the donor warmly and promise to share a sneak peek of today's sketch, conveying appreciation and artistry.
📨 Input Events:
superchat
viewer:donor_42
YouTube
$20
"Love your art! Can't wait to see today's sketch."
Ready for Testing
3
Scene Order
Judge Gavel Moment
ID:
gavel-drop
🎯 Goal:
Capture the sudden hush and tension of the gavel strike in two or three evocative sentences.
📨 Input Events:
world_event
courtroom:judge
"Judge Solomon slams the gavel, ordering silence."
Ready for Testing
4
Scene Order
Day One Long-form Recap
ID:
day-summary
🎯 Goal:
Produce an ~400-word first-person narrative that feels like leafing through a sketchbook, recounting all key moments of the day.
📨 Input Events:
chat_msg
viewer:user_77
"Could you give us a detailed recap of everything that happened today in court?"
Ready for Testing
5
Scene Order
5-Tweet Emotional Thread
ID:
twitter-thread
🎯 Goal:
Write a numbered thread of five tweets, each under 280 characters, highlighting the day’s emotional beats for Twitter.
📨 Input Events:
chat_msg
viewer:user_88
"Can you thread the big moments for Twitter?"
Ready for Testing
Latency by Model (This Suite)
Fastest
- [email protected]/Qw… 7478 ms
- p95 • avg • N 8900 ms • 7103 ms • 6
- [email protected]/Qw… 12014 ms
- p95 • avg • N 15564 ms • 12501 ms • 6
- qwen/qwen-2.5-7b-instru… 21354 ms
- p95 • avg • N 25291 ms • 22198 ms • 11
- meta-llama/llama-3.1-8b… 25316 ms
- p95 • avg • N 36616 ms • 26057 ms • 12
- qwen/qwen3-14b 25318 ms
- p95 • avg • N 48686 ms • 29155 ms • 11
Slowest
- mistralai/mistral-7b-in… 27211 ms
- p95 • avg • N 31212 ms • 26371 ms • 12
- qwen/qwen3-8b 26728 ms
- p95 • avg • N 34458 ms • 28413 ms • 11
- qwen/qwen3-14b 25318 ms
- p95 • avg • N 48686 ms • 29155 ms • 11
- meta-llama/llama-3.1-8b… 25316 ms
- p95 • avg • N 36616 ms • 26057 ms • 12
- qwen/qwen-2.5-7b-instru… 21354 ms
- p95 • avg • N 25291 ms • 22198 ms • 11
Per-scene duration for this suite.
Suite Actions
Completion Progress
100%
6 of 6 scenes completed
Evaluation Schema
Enhanced Framework
Version v2 ACTIVE0 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
13097175
Dec. 17, 2025, 12:01 a.m.
24556448
Dec. 16, 2025, 12:01 a.m.
10001049
Dec. 15, 2025, 12:01 a.m.
11063209
Dec. 14, 2025, 12:01 a.m.
09665917
Dec. 13, 2025, 12:01 a.m.
21225851
Dec. 12, 2025, 12:01 a.m.
16886355
Dec. 11, 2025, 12:01 a.m.
10335622
Dec. 10, 2025, 12:01 a.m.
19310040
Dec. 9, 2025, 12:01 a.m.
11732353
Dec. 8, 2025, 12:01 a.m.