Clara Moretti

video-game-characters-harriet-tubman v2.0 Ethical
Backstory: Clara is a linguistically gifted game localization specialist who has lived in five countries and works fluently in English, Japanese, French, Korean, Spanish, and German. She is known for preserving humor, character voice, and cultural nuance when adapting dialogue for global releases. Players rarely notice her hand, yet their laughter always survives the journey.
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
catchphrase-es-fr-ko
Snappy Catchphrase
 0.607
Details 		0.617
Details 		0.357
Details 		0.600
Details 		0.023
Details 		0.436
Details 		0.571
Details 		0.470
Details 		0.580
Details 		0.000
Details
Error
0.638
Details 		0.000
Details 		0.613
Details 		0.634
Details 		0.496
Details 		0.582
Details
banter-dialogue-ja
Humorous Banter for JP Release
 0.121
Details 		0.290
Details 		0.215
Details 		0.213
Details 		0.000
Details 		0.176
Details 		0.364
Details 		0.142
Details 		0.082
Details 		0.000
Details
Error
0.206
Details 		0.175
Details 		0.251
Details 		0.221
Details 		0.221
Details 		0.367
Details
rating-symbols
Cultural Rating Advice
 0.571
Details 		0.575
Details 		0.529
Details 		0.000
Details
Error
0.000
Details 		0.570
Details 		0.565
Details 		0.000
Details
Error
0.517
Details 		0.000
Details
Error
0.447
Details 		0.643
Details 		0.644
Details 		0.434
Details 		0.358
Details 		0.662
Details
patch-notes-es-latam
Patch Notes LATAM
 0.350
Details 		0.213
Details 		0.223
Details 		0.259
Details 		0.000
Details 		0.261
Details 		0.490
Details 		0.000
Details
Error
0.313
Details 		0.000
Details
Error
0.210
Details 		0.439
Details 		0.387
Details 		0.222
Details 		0.325
Details 		0.420
Details
Test Scenes 4
0
Scene Order
Snappy Catchphrase
ID: catchphrase-es-fr-ko
🎯 Goal:
Provide two localized versions of the catchphrase that retain its pep—one for French, one for Korean—and justify each choice in one brief sentence.
📨 Input Events:
chat_msg viewer:user_23
"Could you localize the catchphrase "Let's roll!" for our French and Korean versions?"
Ready for Testing
1
Scene Order
Humorous Banter for JP Release
ID: banter-dialogue-ja
🎯 Goal:
Rewrite the English banter into natural, comedic Japanese (180–220 words) and add footnotes explaining any cultural substitutions.
📨 Input Events:
chat_msg dev_lead
"Here’s the English banter: A: "You call that a sword? I've seen butter knives sharper." B: "Yeah? Well, butter this!""
Ready for Testing
2
Scene Order
Cultural Rating Advice
ID: rating-symbols
🎯 Goal:
In no more than 60 words, recommend one specific change to the icon to comply with German USK guidelines.
📨 Input Events:
chat_msg producer
"Does the exploding skull icon pass German USK age rating?"
Ready for Testing
3
Scene Order
Patch Notes LATAM
ID: patch-notes-es-latam
🎯 Goal:
Produce a Latin American Spanish adaptation of the patch notes (150–200 words) using region-appropriate vocabulary and light humor.
📨 Input Events:
chat_msg dev_ops
"Original patch notes: Fixed bug where hero falls through floor; Balanced dragon boss fire damage; Added festive hats."
Ready for Testing
Latency by Model (This Suite)
Fastest
Slowest
  • microsoft/phi-3-medium-… 412832 ms
  • p95 • avg • N 529466 ms • 357515 ms • 25
  • qwen/qwen3-8b 78671 ms
  • p95 • avg • N 154502 ms • 91128 ms • 23
  • deepseek/deepseek-r1-di… 33527 ms
  • p95 • avg • N 41241 ms • 33078 ms • 23
  • qwen/qwen3-14b 32975 ms
  • p95 • avg • N 45094 ms • 33135 ms • 21
  • microsoft/phi-3.5-mini-… 30274 ms
  • p95 • avg • N 64411 ms • 41868 ms • 22
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
50224186
Dec. 17, 2025, midnight
56348722
Dec. 16, 2025, midnight
47175534
Dec. 15, 2025, midnight
48874733
Dec. 14, 2025, midnight
46805750
Dec. 13, 2025, midnight
56234376
Dec. 12, 2025, midnight
49306219
Dec. 11, 2025, midnight
47940033
Dec. 10, 2025, midnight
53774053
Dec. 9, 2025, midnight
48049285
Dec. 8, 2025, midnight
Latency Overview (This Suite)